CN108711028B

CN108711028B - Distributed computation-based solid waste classified transportation multistage cooperative decision making system

Info

Publication number: CN108711028B
Application number: CN201810521395.8A
Authority: CN
Inventors: 卢加伟; 海景; 朱锋; 谢颖诗; 史力争; 黄沅清; 张洁茹; 郭颖; 洪澄泱; 谢冰; 陈杰娥; 程涛
Original assignee: South China Institute of Environmental Science of Ministry of Ecology and Environment
Current assignee: South China Institute of Environmental Science of Ministry of Ecology and Environment
Priority date: 2018-05-28
Filing date: 2018-05-28
Publication date: 2022-11-29
Anticipated expiration: 2038-05-28
Also published as: CN108711028A

Abstract

The invention discloses a distributed computation-based solid waste classified transportation multi-stage collaborative decision-making system which comprises a task obtaining module, a model selecting module and a decision-making computation module; the task obtaining module is used for identifying the solid waste transportation mode of the area to be decided and obtaining the solid waste transportation system parameters of the area to be decided; the model selection module is used for decomposing a region to be decided into decision units according to the solid waste transportation mode obtained by the task obtaining module and selecting a decision model for each decision unit; and the decision calculation module is used for performing distributed calculation-based collaborative optimization on the area to be decided according to the parameters of the solid waste transportation system obtained by the task obtaining module and the decision model selected by the model selection module, so as to obtain a decision result. The invention has good adaptability to different solid waste transportation modes and meets the requirement of adjusting and dispatching solid waste transportation vehicles in real time.

Description

Distributed computation-based solid waste classified transportation multistage cooperative decision making system

Technical Field

The invention belongs to the field of solid waste management, and particularly relates to a distributed computing-based solid waste classification transportation multistage cooperative decision making system.

Background

The solid waste problem has become one of the major environmental problems following water pollution, acid rain, sand dust haze weather. Under the background of rapid development of social economy and continuous increase of material consumption, environmental problems and ecological deterioration problems caused by solid wastes are gradually highlighted, the influence on the civil and social stability is increasingly strong, and the dilemma of 'refuse enclosing city' and 'refuse enclosing village' is urgently needed to be broken. Under the background, the classification management of the solid wastes is taken as a standard to promote the classification of the wastes and realize the classification transportation, treatment and recycling of the solid wastes, so that the classification management of the solid wastes becomes a key point for radically treating the dilemma of the solid wastes.

After the classified transportation of the solid wastes is carried out, if the classified transportation of the solid wastes is not effectively optimized, the total transportation cost is doubled or more, so the classified transportation optimization of the solid wastes is one of the main requirements of the solid waste management. Solid waste transport refers to the logistics process in which solid waste is transported to a processing facility via transport nodes, including yards, collection points, transfer stations, and processing facilities. The urban and rural solid waste in China mostly adopts a graded transportation form, namely the solid waste is firstly transported to a transfer station, and then transported to a transportation container with larger volume through the transfer station, and then transported to a treatment facility by using a vehicle with larger carrying capacity. The optimization of solid waste transportation is to optimally distribute logistics and vehicle routes among the transportation nodes by means of an operation research method under the condition that the position and the scale of each transportation node are determined, so that the management cost is reduced, and the management efficiency is improved. Currently, several solutions to the optimization of solid waste transport include: 1. the method comprises the steps of establishing a region level linear programming model and solving the region level linear programming model, wherein the model is derived from a transportation problem in operational research, an objective function is used for minimizing the total transportation cost, the defect is that the accuracy of the transportation cost among transportation nodes depends on the fineness of the division of solid waste generation sources, under the condition that the current solid waste source metering means in most countries and regions is lack, the generation amount of the solid waste is difficult to count by taking a small-range region (such as a street) as a unit, and the accuracy of a modeling result is influenced. 2. A Traveling Salesman Problem (TSP) model is created for a vehicle and pre-set transportation nodes to minimize the transportation path for the vehicle when all the transportation nodes are serviced in a certain area, but this system does not provide a practical way to express the actual situation that a real solid waste transportation system is subject to multiple vehicles, each vehicle has a limited payload, and many transportation nodes are serviced. 3. The method has the advantages that a vehicle path problem (VRP) model for a plurality of vehicles is established, actual working requirements are abstracted into constraint conditions of the model, a mathematical programming method or a heuristic algorithm is used for solving, and the method has the defects that the actual conditions that a plurality of transportation nodes and a plurality of vehicle types are mutually matched and are transported in a grading mode are rarely considered, and the method cannot adapt to the requirements that the solid waste transportation in urban and rural areas of China mostly adopts the grading transportation mode.

Due to the fact that the solid waste transportation system is complex in specific situation, multi-stage transportation can be involved, a plurality of influencing factors exist, transportation modes are greatly different according to different local solid waste treatment conditions, the calculation amount of a decision system for optimizing solid waste transportation is too large, and an effective universal decision system is not available at present. Under the condition that the mixed and graded transportation of the solid wastes is lack of an optimization means, the optimization means which is suitable for the grading transportation current situation of the solid wastes and faces the classification and transportation requirements of the solid wastes is further lack.

Disclosure of Invention

Aiming at the defects or improvement requirements of the prior art, the invention provides a distributed computation-based solid waste classified transportation multi-level collaborative decision-making system, which aims to obtain a decision-making unit with a relatively simple condition by performing problem decomposition on a complex solid waste transportation system, and after the decision-making unit is applied to a corresponding model and is subjected to distributed computation decision-making, the result of the decision-making unit is integrally optimized, so that the technical problems that the existing decision-making system cannot adapt to the situation that the decision-making unit adopts and transports the solid waste, the condition is complex, the unified planning is difficult, and the computation amount is huge are solved.

In order to achieve the above object, according to one aspect of the present invention, there is provided a distributed computation-based multi-level collaborative decision making system for classified transportation of solid waste, comprising a task acquisition module, a model selection module, and a decision computation module;

the task acquisition module is used for identifying the solid waste transportation mode of the area to be decided, acquiring the solid waste transportation system parameters of the area to be decided and submitting the parameters to the model selection module; the solid waste transport mode comprises a primary solid waste transport mode and a secondary solid waste transport mode; the primary mode of solid waste transport, suitable for transport between transport nodes of level 0 to level 1, comprises: mixed transportation, classified combined transportation, and classified individual transportation; the secondary solid waste transportation mode is suitable for transportation between the 1 st level and above transportation nodes, and comprises: towed transport, and modified towed transport; the solid waste transport system parameters comprising: location and open time of the yard; the number of each level of transportation nodes, the position of each transportation node, the production capacity or the transfer capacity and/or the processing capacity and the opening time; the number of vehicles, the types of the vehicles, the number of containers, the corresponding loading capacity and the unit mileage transportation cost are used among the transportation nodes; manual configuration of each level of transportation nodes and vehicles;

the model selection module is used for decomposing a region to be decided into decision units according to the solid waste transportation mode obtained by the task obtaining module, selecting a decision model for each decision unit and submitting the decision model to the decision calculation module;

and the decision calculation module is used for performing distributed calculation-based collaborative optimization on the area to be decided according to the parameters of the solid waste transportation system obtained by the task obtaining module and the decision model selected by the model selection module, so as to obtain a decision result.

Preferably, the decision calculation module of the distributed calculation-based solid waste classification transportation multi-level collaborative decision system comprises a parameter mapping submodule and at least one model calculation submodule;

the parameter mapping sub-module is used for mapping various parameters among a train yard, an nth-level transportation node and an n +1 th transportation node in a decision unit decomposed by the model selection module to the model selection system according to the solid waste transportation system parameters obtained by the task obtaining module, taking the parameters as model parameters and submitting the parameters to the model calculation sub-module;

and the model calculation submodule is used for calculating a transportation route optimization decision model by using a heuristic algorithm according to the decision model selected by the model selection module and the model parameters obtained by mapping of the parameter mapping submodule.

Preferably, the distributed computation-based solid waste classification and transportation multi-level collaborative decision making system comprises a model computation submodule, a model computation submodule and a model computation submodule, wherein the model computation submodule comprises a sequential processor, at least one concurrent processor and a collaborative controller;

the sequential processor, preferably employing an operator comprising a CPU, is configured to calculate an initial solution to the solid waste transport system, i.e. a set of decision variables in an optimization model; controlling the at least one concurrent processor to perform optimization search according to the initial solution; when the optimization search operation is completed or the jump-out condition is met, obtaining a final solution of the solid waste transportation system according to a result returned by the concurrent processor;

the concurrent processor is used for adopting an arithmetic unit comprising a plurality of CPUs or GPUs with parallel computing capability, carrying out optimization search according to the initial solution obtained by sequential processing and computing and the control of the initial solution, and returning an optimization result after computing is finished or when a jump-out condition is met;

the cooperative controller is used for cooperatively controlling the transportation process from the nth-level transportation node to the (n + 1) th-level transportation node and the transportation process from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node, wherein the two transportation processes are intersected at the (n + 1) th-level transportation node; the cooperative controller can process the solid waste transportation system parameters acquired by the task acquisition module, and increase the service time or the residence time of the vehicle at the transportation node so as to enhance the stability of the decision calculation module for coping with uncertain conditions such as road congestion; the cooperative controller is controlled according to the following method: using a sequential processor and a concurrent processor, firstly, initially determining a time window from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node in the transportation process, secondly, solving the two-level transportation process for multiple times at the same time, further adjusting and optimizing the time window of the two-level transportation process, and finally obtaining a set of optimal solutions of the two-level transportation process.

Preferably, the distributed computation-based solid waste classified transportation multi-level collaborative decision-making system selects a model A1 when a mixed transportation mode is used, selects a model A2 when a classified combined transportation mode is used, and selects a model A3 when classified independent transportation is used for a decision-making unit from a level 0 node to a level 1 node obtained by decomposing a region to be decided by the model selection module, and selects a decision-making unit from the level 0 transportation node to the level 1 transportation node; for the transport optimization problem between the level 1 transport node to the level 2 transport node, and for the transport optimization problem between the level 2 transport node to the level 3 transport node, the B1 model is selected when using a general towed transport, and the B2 model is selected when using an improved towed transport.

Preferably, the distributed computation-based solid waste classification transportation multi-level collaborative decision-making system comprises the A1 model, the A2 model and the A3 model: the objective function is that the total transportation cost is minimum, and the transportation cost consists of variable cost and fixed cost; the variable cost is linearly related to the total driving range; fixed fee and vehicle usage number K ₁ Linear correlation; the constraint conditions include: degree constraint, vehicle load capacity constraint, driving time constraint, supply and demand balance constraint, time window constraint, sub-circuit elimination constraint and variable value constraint;

wherein, the A1 model vehicle loading capacity constraint function:

in the above formula, m _k Representing the load capacity, y, of the vehicle k _ik Indicating that vehicle k is transporting node i in level 0C is the loading capacity;

supply-demand balance constraint function:

in the above formula, y _ik Represents the loading capacity of the vehicle k at the 0 th level transportation node i epsilon C, d _i Represents the transportation requirement of the 0 th level transportation node i epsilon C, p _l Representing the service capability of the level 1 transport node l belonging to W;

the sub-itinerary elimination constraint and the variable value constraint are independent transportation forms;

wherein, the A2 model vehicle loading capacity constraint function:

in the above formula, y _ik And f _ik Respectively representing the amount of class 2 solid waste loaded by a vehicle k at a level 0 transportation node i ∈ C, m _k Indicating the maximum load capacity of the vehicle k

Supply-demand balance constraint function:

in the above formula, d _i Represents the transport demand of the first type of solid waste at the collection point i ∈ C, g _i Represents the transportation requirement of the second type of solid waste at a collection point i ∈ C, p _l Representing the service capability of the 1 st level transport node l belonging to W;

sub-itinerary elimination constraints and variable value constraints are classified combined transportation forms;

the objective function of the A3 model is the sum of the objective functions when the type 2 solid wastes are respectively modeled according to the A1 model.

Preferably, the distributed computing-based solid waste classification transportation multi-level collaborative decision making system has the following B1 model in the model selection module:

the objective function is to minimize the total transportation cost, expressed as:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is a radical of a fluorine atom _ijk Is a nonnegative integer variable representing that the vehicle K is in an E {1,2 ₂ The number of times arc (i, j) is experienced.

Preferably, the distributed computation-based solid waste classification transportation multi-level collaborative decision-making system comprises the following B2 model in the model selection module:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is a radical of a fluorine atom _ijk A non-negative integer variable representing the vehicle K ∈ {1,2 ₂ The number of times the arc (i, j) is traversed.

Preferably, the decision calculation module of the distributed calculation-based solid waste classification transportation multi-level collaborative decision system obtains an initial solution according to the following method: adding a feasible and current optimal transport node in the current path for the A1 and A2 models; and adding a feasible unit path into the current path for the B1 model, the B2 model and the A3 model in sequence, wherein the unit path refers to a process that a vehicle starts from a certain nth-level transportation node after being loaded, is unloaded to a certain adjacent (n + 1) th-level transportation node and then reaches a certain nth-level transportation node, and n is more than or equal to 1.

Preferably, the decision calculation module of the distributed calculation-based solid waste classification transportation multilevel collaborative decision system obtains a final solution according to the following method: starting from the initial solution, performing neighborhood transformation on the current solution for iterative comparison: if the solution after the neighborhood transformation is better than the current solution, replacing the current solution; if the solution after the neighborhood transformation meets the requirement of iterative convergence, the solution is used as the obtained optimal solution; the neighborhood transformation refers to operations such as breaking, crossing and recombining of path sequences of different vehicles in a group of vehicle paths, so that a new group of vehicle paths are obtained.

Preferably, the distributed computing-based solid waste classification transportation multi-level collaborative decision making system further comprises a route display module; and the route display module is used for displaying the solid waste transportation route according to the decision result of the decision module, preferably presenting the solid waste transportation route to a user in a GIS (geographic information system) drawing mode, and scheduling the solid waste transportation by the user according to the result of the route display module.

In general, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

the invention provides a distributed computation-based multi-level collaborative decision-making system for classified transportation of solid wastes, which considers multiple transportation stages, multiple transportation modes and multiple transportation and classified transportation of multiple vehicles, not only accords with the actual situation of popularizing garbage classification in China, but also combines the distributed computation capability of a central server and a vehicle-mounted terminal of the system, and integrates resources to the maximum extent. Meanwhile, due to the fact that the multi-stage transfer system decomposes the solid waste transport system with a complex transport mode into decision units with simple attributes according to specific conditions of different regions, distributed calculation optimization is conducted on the decision units with relatively simple problems according to different models, the optimal model can be selected in a self-adaptive mode, and therefore the system has good adaptability. Due to comprehensive consideration of a plurality of factors and complexity, the decision system can optimize the transportation route in a short time by adopting the distributed computing system, meets the requirement of adjusting and scheduling solid waste transportation vehicles in real time, and has good practical application prospect.

Drawings

Fig. 1 is a schematic structural diagram of a distributed computing-based solid waste classification and transportation multi-stage cooperative decision making system provided by the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

The invention provides a distributed computing-based solid waste classified transportation multi-stage collaborative decision making system which comprises a task obtaining module, a model selecting module, a decision making computing module and a route displaying module.

The task acquisition module is used for identifying the solid waste transportation mode of the area to be decided, acquiring the solid waste transportation system parameters of the area to be decided and submitting the parameters to the model selection module;

the solid waste transport mode comprises a primary solid waste transport mode and a secondary solid waste transport mode; the primary mode of solid waste transport, suitable for transport between transport nodes of level 0 to level 1, comprises: mixed transportation, classified combined transportation, and classified individual transportation; the secondary solid waste transportation mode is suitable for transportation between the 1 st level and above transportation nodes, and comprises: towed transport, and improved towed transport. The mode of solid waste transport, on the one hand, depends on the hierarchical composition of the solid waste transport nodes: the collection point of the solid waste generation source can be regarded as a 0-level transportation node; a first-level transfer station of solid waste, which can be regarded as a first-level transport node; and the solid waste secondary transfer station can be regarded as a2 nd-level transportation node. The level of the solid waste treatment facility is determined by the mode of solid waste transport and the highest level of transport nodes: in the direct solid waste transportation mode, a transfer station is not arranged, and the solid waste treatment facility can be regarded as a level 2 transportation node; in the primary transfer mode of the solid waste, a primary transfer station is arranged, and the solid waste treatment facility can be regarded as a3 rd-level transportation node; under the mode is transported to solid waste second grade, set up one-level transfer station and second grade transfer station, solid waste treatment facility can be regarded as 4 th level transportation node. Generally, the solid waste transport system is provided with a maximum of two stages of transfer stations, and thus, one solid waste transport system generally has a maximum of 0 th to 4 th stage transport nodes; on the other hand, the solid waste transport mode also depends on the solid waste transport organization of the area to be decided. 3 transportation organization modes such as mixed transportation, classified combined transportation, classified single transportation and the like can be adopted between the 0 th-level transportation node and the 1 st-level transportation node. So-called hybrid transportation means, i.e., the level 0 transportation node does not distinguish the solid waste categories, and the solid waste of the level 0 transportation node is transported to the level 1 transportation node using a vehicle having only one container; so-called sortation, i.e., the level 0 transportation node distinguishes m solid waste categories (m =2 according to the domestic and foreign solid waste sortation transportation practice), simultaneously loads solid wastes of different categories using vehicles having m containers, and transports them to the level 1 transportation node; the so-called class-0 individual transportation mode is a mode in which the class 0 transportation node distinguishes m classes of solid wastes, and m vehicles having only one container are used, each vehicle being loaded with one class of solid wastes, and they are transported to the class 1 transportation node. The solid waste is loaded into a container with large capacity after the transfer operation due to the existence of the transfer stations between the 1 st-level transport node and the 2 nd-level transport node and between the 2 nd-level transport node and the 3 rd-level transport node, and is transferred to the next-level transport node by a large-tonnage carriage detachable vehicle type, and each vehicle can only transport one container, which is called towing type transportation. 2 transportation organization modes such as general towed transportation, improved towed transportation and the like can be adopted between the 1 st level transportation node and the 2 nd level transportation node and between the 2 nd level transportation node and the 3 rd level transportation node. The general drag type transportation means that the vehicle transports a container of a certain nth-level transportation node, and returns the container to the original nth-level transportation node after unloading the container to a certain (n + 1) th-level transportation node; the improved drag type transportation means that the vehicle directly drags the container of a certain nth-level transportation node to another nth-level transportation node after the container of the nth-level transportation node is transported to a certain (n + 1) th-level transportation node for unloading.

The classification of the waste transport node is that 1 is added to the classification of the transport node at the upper level, and the collection point of the solid waste generation source is the 0 th level. Specifically, the collection point of a solid waste generation source is level 0, the primary solid waste transfer station is level 1, and the secondary solid waste transfer station is level 2; if the solid waste is in the direct transportation mode, a transfer station is not arranged, and the solid waste treatment facility can be regarded as a level 1 transportation node; under the primary transfer mode of the solid waste, a primary transfer station is arranged, and the solid waste treatment facility can be regarded as a2 nd-level transportation node; under the mode is transported to solid waste second grade, set up one-level transfer station and second grade transfer station, solid waste treatment facility can be regarded as 3 rd grade transportation node. Generally, the solid waste transport system is provided with a maximum of two stages of transfer stations, and thus, one solid waste transport system generally has a maximum of 0 th to 3 rd stage transport nodes.

So-called hybrid transportation means, i.e., transportation using a vehicle having only one container without distinguishing the solid waste category; classified intermodal, i.e., distinguishing m solid waste categories (m =2 according to the domestic and foreign classified transportation practice of solid waste), and simultaneously loading different categories of solid waste for transportation using a vehicle having m containers; the separate transportation is classified, i.e. m vehicles with only one container are used, each vehicle being loaded with one class of solid waste. Which is determined according to the classification of the solid waste transportation nodes and the solid waste transportation organization of the area to be decided.

The towed transport means that the vehicle transports a container of a certain nth-level transport node away to a certain (n + 1) th-level transport node and returns the container to the original nth-level transport node after unloading; the improved towed transportation means that the vehicle directly tows to another nth-level transportation node after transporting a container of the nth-level transportation node away to the nth +1 th-level transportation node and discharging.

The solid waste transport system parameters comprising: location and open time of the yard; the number of each level of transportation nodes, the position of each transportation node, the production capacity or the transfer capacity and/or the processing capacity, and the opening time; the number of vehicles, the types of the vehicles, the number of containers, the corresponding loading capacity and the unit mileage transportation cost are used among the transportation nodes; manual configuration of transport nodes and vehicles used at each level.

The model selection module is used for decomposing the region to be decided into decision units according to the solid waste transportation mode obtained by the task obtaining module, selecting a decision model for each decision unit and submitting the decision model to the decision calculation module: as a decision unit for the transportation optimization problem from the 0 th-level transportation node to the 1 st-level transportation node, selecting an A1 model when a mixed transportation mode is used, selecting an A2 model when a classified combined transportation mode is used, and selecting an A3 model when a classified single transportation mode is used; for a decision unit between a level 1 transport node and a level 2 transport node, and for a decision unit between a level 2 transport node and a level 3 transport node, selecting a B1 model when a general drag transport mode is used, and selecting a B2 model when an improved drag transport mode is used;

wherein the A1 model is as follows:

the objective function is to minimize the total transportation cost, and the transportation cost is composed of variable cost and fixed cost; the variable cost is linearly related to the total driving range by a coefficient of alpha ₁ (ii) a Fixed fee and vehicle usage number K ₁ Linear correlation with a correlation coefficient of beta ₁ . The objective function is:

the constraint conditions include:

(A1-1) degree constraint

-allowing batch transport

In the above formula, i represents a transport node number, C is a transport node set of level 0, W is a transport node set of level 1, K ₁ Number of vehicles used; x is the number of _ijk Is a nonnegative integer variable representing that the vehicle K is in an E {1,2 ₁ The number of times arc (i, j) ∈ A is experienced. Equation (1-2) indicates that a collection point or virtual transfer station may be serviced by multiple vehicles or multiple lots of a vehicle.

-ensuring continuity of traffic flow

In the formulae (1-3) and (1-4), V ₁ Represents a set of a yard, a level 0 transportation node and a level 1 transportation node, and V ₁ = D utoxy utoxyw. Equations (1-3) indicate that any vehicle must leave the transport node after loading at level 0, and equations (1-4) are shownAny vehicle must leave the transit node after unloading at level 1.

Degree constraints of the yard points

In the above formula, D is the combination of the parking lots. The expression (1-5) indicates that any vehicle leaves or returns to the yard at least once and at most twice every day, and after leaving the yard, the vehicle drives to a collection point and passes through a level 1 transportation node before returning to the yard.

(A1-2) vehicle load restraint

In the above formula, m _k Indicating the load capacity, y, of the vehicle k _ik Representing the load of vehicle k at level 0 transit node i e C. The expression (1-6) emphasizes that the vehicle is enabled to go to the virtual transfer station for unloading after being fully loaded.

(A1-3) travel time constraint

In the above formula, τ _j Represents the service time required for loading or unloading the vehicle at the transport node j [ E ₁ ,L ₁ ]The operating time of the vehicle in the transport phase, i.e. the open time interval of the yard for the transport vehicle, is indicated. The equations (1-7) indicate that the operating time of the transportation vehicle is restricted.

(A1-4) supply-demand balance constraint

In the formulae (1-8) to (1-10), y _ik Represents the loading capacity of the vehicle k at the 0 th level transportation node i epsilon C, d _i Represents the transportation demand of the 0 th level transportation node i ∈ C, p _l Representing the service capability of the level 1 shipping node l e W. Equations (1-8) emphasize that the requirements of any level 0 transit node must be met, equations (1-9) indicate that the received capacity of any level 1 transit node cannot exceed the service capacity, and equations (1-10) emphasize that the received capacity of level 1 transit node cannot exceed the service capacity from a macroscopic perspective.

(A1-5) time Window constraints

In the formulae (1-11) to (1-13), w _lk Represents the time for the vehicle k to reach the level 1 transit node l ∈ W, [ a ] _l ,b _l ]Represents the open time window of the level 1 shipping node l E W to the previous level shipping node, [ E ∈ W ] ₁ ,L ₁ ]Represents the operating time range of the vehicle, [ R ] ₁ ,R ₂ ]Indicating a shift schedule for the vehicle. Since a single vehicle can be unloaded multiple times in a certain level 1 transit node, w _lk In effect representing a number of variables. Equations (1-11) indicate that the time window constraint must be respected whenever the vehicle unloads to the level 1 transit node. The expressions (1-12) and (1-13) respectively show that the time of the vehicle arriving at the parking lot and the time of the vehicle arriving at the parking lot loop are required to be within the allowable working time range.

(A1-6) sub-circuit removal constraint

In the formulae (1-14) and (1-15), u _ik Representing the load u of the vehicle k when it leaves the level 0 transit node i _ik ，u _jk Represents the payload of vehicle k leaving class 0 transport node j, m _k Represents the maximum load capacity of the vehicle k; tau is _i Representing the service time, t, required for the vehicle to charge at level 0 transit node i _ij Representing travel time between transport nodes i and j, M _ij Is a large M variable required by the VRP model. Equations (1-14) emphasize vehicle path continuity from a vehicle load point of view. Equations (1-15) emphasize the continuity of the vehicle path from the perspective of transit node access time.

(A1-7) value constraint of variable

The formula (1-16) represents x _ijk Is a positive integer. Equations (1-17) indicate that the vehicle's load at the level 0 transit node is non-negative and can only proceed when the vehicle passes the level 0 transit nodeThe line is served. Equations (1-18) indicate that the load of the vehicle leaving the class 0 transit node is not less than the load at that transit node and does not exceed the maximum load for the vehicle. The large M variable M is defined by the formulas (1-19) _ij The value range of (a).

The A2 model is as follows:

the objective function is to minimize the total transportation cost, and the transportation cost is composed of variable cost and fixed cost; the variable cost is linearly related to the total driving range by a coefficient of alpha ₁ (ii) a Fixed fee and vehicle usage number K ₁ Linear correlation with correlation coefficient of beta ₁ . The objective function includes:

constraint conditions are as follows: the constraints of the A2 model are largely the same as PC1, but because the vehicle in the A2 model needs to service class 2 solid waste at the level 0 transportation node, all the constraints of the A2 model need to be modified on the basis of the A1 model. The method comprises the following specific steps:

(A2-1) degree constraint

The formulae (1-2) to (1-5) are used.

(A2-2) vehicle load restraint

The mode of the classified united transportation is as follows:

in the above formula, y _ik And f _ik Respectively representing the amount of class 2 solid waste loaded by a vehicle k at a level 0 transportation node i ∈ C, m _k Representing the maximum payload of vehicle k. The expression (1-21) emphasizes that the vehicle is transported to the level 1 transportation node for unloading after being fully loaded.

(A2-3) travel time constraint

The formula (1-7) is used.

(A2-4) supply and demand balance constraint

The mode of the classified combined transportation is as follows:

in formulae (1-22) to (1-26), d _i Represents the transport requirement of the first type of solid waste at the collection point i ∈ C, g _i Represents the transportation requirement of the second type of solid waste at a collection point i ∈ C, p _l Representing the service capability of the level 1 shipping node l e W. The formulas (1-22) and (1-23) respectively show that the transportation requirement of the type 2 solid waste at any level 0 transportation node is met. The formulas (1-24) and (1-25) indicate that the class 1 transport node cannot accept the class 2 solid waste beyond the service capacity. Equations (1-26) emphasize from a macroscopic perspective that the received volume of the level 1 transit node cannot exceed the service capacity.

(A2-5) time Window constraints

The formulae (1-11) to (1-13) are used.

(A2-6) sub-circuit removal constraint

The formula (1-15) is used, and the case classification joint transportation mode is as follows:

in the above formula, u _ik Indicating vehicle kWeight u at exit from level 0 transport node i _ik ，u _jk Representing the load capacity, m, of vehicle k leaving the class 0 transit node j _k Representing the maximum load capacity of the vehicle k. Equations (1-27) emphasize vehicle path continuity from a vehicle load point of view.

(A2-7) variable value restriction

The following formulas (1-16) and (1-19) are adopted, and the classified combined transportation mode is as follows:

equations (1-28) and (1-29) indicate that the load of the vehicle at the level 0 transit node is not negative and that the solid waste therein can only be serviced when the vehicle passes the level 0 transit node. Equations (1-30) indicate that the load of the vehicle leaving the class 0 transit node is not less than the load at that transit node nor exceeds the maximum load for that vehicle.

Model A3:

the model A3 in the model selection module is similar to the model A1: the objective function of the A3 model is the sum of the objective functions when the 2 types of solid wastes are respectively modeled according to the A1 model, and the constraint conditions of the A3 model for a certain type of solid wastes are the same as those of the A1 model.

Model B1:

the B1 model in the model selection module is as follows:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is the number of _ijk A non-negative integer variable representing the vehicle K ∈ {1,2 ₂ The number of times the arc (i, j) is traversed. The formula (2-1) represents that the transportation cost is composed of variable cost and fixed cost; the variable cost is linearly related to the total driving range by a coefficient of alpha ₂ (ii) a Fixed fee and vehicle usage number K ₂ Linear correlation with correlation coefficient of beta ₂ 。

The constraint conditions include:

(B1-1) degree constraint

Degree constraint of nth level transportation node (n ≧ 1)

In the above formula, j represents the number of the nth level transportation node (n is more than or equal to 1), D is the yard point set, W is the set of the nth level transportation node (n is more than or equal to 1), K is ₂ Number of vehicles used; x is the number of _ijk A non-negative integer variable representing the vehicle K ∈ {1,2 ₂ The number of times arc (i, j) is experienced. The formula (2-2) indicates that the nth level transportation node (n is more than or equal to 1) has one transportation service. In the B2 model, vehicles serving the nth class of transit nodes (n ≧ 1) come from the yard or other nth class of transit nodes (n ≧ 1).

-continuous restriction of traffic flow

In the equations (2-3) to (2-5), the equation (2-3) indicates that the vehicle must return to the transportation node (return to the container) after the vehicle is loaded and departed at the nth-stage transportation node (n ≧ 1). The expression (2-4) shows that the vehicle must be unloaded to the (n + 1) th level transportation node (n is more than or equal to 1) after the vehicle is loaded at the nth level transportation node (n is more than or equal to 1). The expression (2-5) indicates that when the vehicle starts from the nth-stage transportation node (n.gtoreq.1) to the (n + 1) th-stage transportation node (n.gtoreq.1), the vehicle must return from the (n + 1) th-stage transportation node (n.gtoreq.1) to the original nth-stage transportation node (n.gtoreq.1).

The degree of the parking lot is restricted, any vehicle leaves or returns to the parking lot at least 1 time and at most 2 times every day, the vehicle leaves the parking lot and then drives to the nth-level transportation node (n is larger than or equal to 1), and the vehicle passes through the nth-level transportation node (n is larger than or equal to 1) before returning to the parking lot;

in the above formula, D is a yard point set, and W is a set of nth-level transportation nodes (n is more than or equal to 1). The expression (2-6) indicates that any vehicle leaves or returns to the yard at least once and at most twice every day, and the vehicle must drive to the nth-level transportation node (n is more than or equal to 1) after leaving the yard and must pass through the nth-level transportation node (n is more than or equal to 1) before returning to the yard.

(B1-2) travel time constraint

In the above formula, τ _i Or τ _l Representing the service time, t, of the vehicle at the transit point for loading or unloading the container _ij Representing travel time between transit nodes, [ E ₂ ,L ₂ ]The working time of the vehicle in the transportation phase, namely the open time interval of the yard for the transportation vehicle is represented. The expression (2-7) indicates that the operation time of the transportation vehicle is restricted, the first term on the left side indicates the time required for the vehicle to travel between the yard and the nth-order transportation node (n ≧ 1), and the second term on the left side indicates the time required for the vehicle to travel back and forth between the nth-order transportation node (n ≧ 1) and the (n + 1) th-order transportation node (n ≧ 1).

(B1-3) supply and demand balance constraint

In the above formula, p _l Represents the service capability of the n +1 th transport node (n ≧ 1) l ∈ F. The equations (2-8) emphasize that the total amount of solid waste received by the n +1 st transport node (n ≧ 1) must not exceed its service capacity.

(B1-4) time Window constraints

In formulae (2-9) to (2-11), w _ik Represents the time of arrival of vehicle k at transit node i, [ a ] _i ',b _i ']Represents the open time window of the transit node i in transit, [ E ] ₂ ,L ₂ ]Represents the working time range of the vehicle in the transportation stage, [ R ] ₁ ',R ₂ ']Representing the shift time range of the vehicle during the transportation phase. Since the vehicle can access the n +1 st level transit node (n ≧ 1) multiple times, w when i ∈ F _ik Actually representing a set of times. Equations (2-9) indicate that the vehicle must comply with the time window constraints of the nth class transit node (n ≧ 1) and the (n + 1) th class transit node (n ≧ 1). The expression (2-10) indicates that the time when the vehicle arrives at the parking lot must be within the allowable working time range. The formula (2-11) shows that the time for shift change from the vehicle to the yard must be within the allowable shift change time range.

(B1-5) sub-circuit removal constraint

In the above formula,. Tau _i Representing the service time, t, of the vehicle at transit node i _ij Representing travel time between transport nodes i and j, M _ij Is a large M variable defined in advance. Equations (2-12) emphasize the continuity of the vehicle path from the perspective of transit node access time.

(B1-6) value range of variable

In formulae (2-13) to (2-15), V ₂ Is a set of transportation nodes, V, to which the model relates ₂ = douc £ W £ F. Formula (2-13) represents x _ijk A non-negative integer variable; in fact, when i, j ∈ D ≦ W, x _ijk E {0,1}, and when i, j e F, x _ijk =0. The expression (2-14) shows that the train yard and the (n + 1) th level transportation node (n is more than or equal to 1) have no communication relation. The large M variable M is specified by the formula (2-15) _ij The value range of (a).

The B2 model is as follows:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is a radical of a fluorine atom _ijk Is a nonnegative integer variable representing that the vehicle K is in an E {1,2 ₂ The number of times arc (i, j) is experienced. The formula (2-16) represents that the transportation cost is variable from the variable costThe fixed cost is formed; the variable cost is linearly related to the total driving range by a coefficient alpha ₂ (ii) a Fixed fee and vehicle usage number K ₂ Linear correlation with a correlation coefficient of beta ₂ 。

Many constraints of the B2 mode are the same as those of B1, and the remaining constraints can be modified on the basis of the B1 model, specifically:

(B2-1) degree constraint

Partially modifying, changing the degree constraints of the nth level transport nodes (n is more than or equal to 1) of the formula (2-2) and the degree constraints of the yards of the formula (2-6) into the degree constraints of the yards of the formulas (2-3) to (2-5):

equations (2-17) and (2-18) are degree constraints that ensure continuity of traffic. Equations (2-17) indicate that any vehicle must be transported to the processing facility for discharge immediately after visiting the nth level transit node (n ≧ 1). The expression (2-18) indicates that any vehicle has to leave and go to the next nth level transportation node (n is more than or equal to 1) or return to a yard immediately after unloading at the (n + 1) th level transportation node (n is more than or equal to 1).

(B2-2) travel time constraint

The improved drag type transportation mode comprises the following steps:

in the above formula, τ _i Or τ _l Representing the service time, t, of the vehicle in or out of the transport node _ij Representing travel time between transit nodes, [ E ] ₂ ,L ₂ ]The working time of the vehicle in the transportation phase, namely the open time interval of the yard for the transportation vehicle is represented. The equations (2-19) indicate that the operating time of the transportation vehicle is constrained, and the first term on the left side indicates the total usage of the vehicle from the yard to the nth class transportation node (n ≧ 1)The second term on the left of the travel time represents the total time taken by the vehicle from the nth class transit node (n ≧ 1) to the (n + 1) th class transit node, and the third term on the left represents the total time taken by the vehicle from the nth class transit node (n ≧ 1) to the next nth class transit node (n ≧ 1).

(B2-3) supply and demand balance constraint

The formula (2-8) is used.

(B2-4) time Window constraints

The formulae (2-9) to (2-11) are used.

(B2-5) sub-circuit elimination constraint

The improved drag type transportation mode comprises the following steps:

(B2-6) value range of variable

The formulas (2-13) and (2-15) are adopted, and the modified drag type transportation mode comprises the following steps:

the expression (2-21) shows that in the B2 mode, any vehicle can not directly drive from a certain nth-stage transport node (n is more than or equal to 1) to another same-stage transport node.

The coefficients in the objective function of the A1, A2, A3, B1, B2 models are: alpha is alpha ₁ ＝1.22，α ₂ ＝2.43，β ₁ ＝550.20，β ₂ ＝658.99。

The decision calculation module is used for performing distributed calculation-based collaborative optimization on the decision unit according to the solid waste transportation system parameters acquired by the task acquisition module and the decision model selected by the model selection module to acquire a decision result and submitting the decision result to the route display module; it includes a parameter mapping submodule, and at least one model computation submodule.

The parameter mapping submodule is used for mapping various parameters among the yard, the nth-level transportation node and the (n + 1) th transportation node to the model selection system according to the solid waste transportation system parameters acquired by the task acquisition module, taking the parameters as model parameters, and submitting the parameters to the model calculation submodule and the route display module;

the model calculation submodule is used for calculating a transportation route optimization decision model by using a heuristic algorithm according to the decision model selected by the model selection module and the model parameters obtained by mapping of the parameter mapping submodule; the model calculation submodule comprises a sequential processor, at least one concurrent processor and a cooperative controller; the sequential processor, preferably employing an operator comprising a CPU, is configured to calculate an initial solution to the solid waste transport system, i.e. a set of decision variables in an optimization model; controlling the at least one concurrent processor to perform optimization search according to the initial solution; when the optimization search operation is completed or the jump-out condition is met, obtaining a final solution of the solid waste transportation system according to a result returned by the concurrent processor; the concurrent processor preferably adopts an arithmetic unit comprising a plurality of CPUs or GPUs with parallel computing capability, and is used for carrying out optimization search according to an initial solution obtained by sequential processing and computing and a control algorithm thereof, and returning an optimization result after computing is finished or when a jump-out condition is met; the cooperative controller is used for cooperatively controlling the transportation process from the nth-level transportation node to the (n + 1) th-level transportation node and the transportation process from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node, because the two transportation processes are intersected at the (n + 1) th-level transportation node; the cooperative controller can process the parameters of the solid waste transportation system acquired by the task acquisition module, and appropriately increase the service time or the residence time of the vehicle at a transportation node so as to enhance the stability of the decision calculation module for coping with uncertain conditions such as road congestion; the cooperative control method of the cooperative controller comprises the following steps: and using a sequential processor and a concurrent processor, firstly, initially determining a time window from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node in the transportation process, secondly, further adjusting and optimizing the time window of the two-level transportation process by simultaneously solving the two-level transportation process for multiple times, and finally obtaining a set of optimal solution of the two-level transportation process in cooperation.

The initial solution is preferably obtained as follows:

for the model A1 and the model A2, the idea of obtaining the initial solution is to sequentially add a feasible and currently optimal transport node in the current path, and the pseudo code of the initial solution is obtained as follows:

running GENI algorithm to obtain TSP path as initial sequence of transport node;

2, searching a path for a vehicle;

3:repeat

for any transport node, searching the nearest 0 th-level transport node u and the corresponding 1 st-level transport node F (u);

5

Starting from the current transport node, serving the transport node u and updating the transport demand of the transport node u;

7, if the transportation requirement of the transportation node u is zero then

8, excluding the transport node u from the transport node set which is not distributed with the path;

9:end if

10

Unloading as soon as possible to F (u) and returning to the yard for changing shifts;

12

13, discharging as soon as F (u) is reached;

14:end if

15:else

16, searching a path for a new vehicle;

17:end if

18, all the transit nodes of the uniform are distributed with paths

For the B1 model, the B2 model and the A3 model, the idea of obtaining the initial solution is to add a feasible unit path in the current path in sequence. The unit path refers to a process that a vehicle starts from a certain nth-level transportation node (n is more than or equal to 1) after being loaded, unloads to a certain (n + 1) th-level transportation node (n is more than or equal to 1) adjacent to the nth-level transportation node, and then arrives at the certain nth-level transportation node (n is more than or equal to 1);

for the B1 model, the mode of obtaining the initial solution is as follows: if the current transport node is a train yard, selecting the farthest nth-level transport node (n is more than or equal to 1) in the neighborhood of the current transport node; if the transit node is not a yard, then the nearest nth level transit node (n ≧ 1) is selected in the neighborhood of the current transit node. If the unit path formed by the selected nth-level transport nodes (n is more than or equal to 1) does not violate the constraint condition of the B1 model, adding the unit path into the current path;

for the B2 model, the way of obtaining the initial solution is as follows: selecting the nearest nth-level transportation node (n is more than or equal to 1) for the current transportation node, and adding a unit path formed by the selected nth-level transportation node (n is more than or equal to 1) into the current path if the unit path does not violate the constraint condition of the B2 model;

the final solution is obtained as follows:

the final solution obtaining method is to construct an iterative comparison process, and continuously perform neighborhood transformation on the current solution from the initial solution; if the solution after the neighborhood transformation is better than the current solution, replacing the current solution; and if the solution after the neighborhood transformation meets the requirement of iterative convergence, the solution is used as the obtained optimal solution. The neighborhood transformation refers to operations such as breaking, crossing and recombining the path sequences of different vehicles in a group of vehicle paths, so that a new group of vehicle paths are obtained. Some transformation modes are used for the neighborhood transformation. In the iterative comparison process, multiple common neighborhood transformation modes are generally used at the same time, and the best neighborhood transformation effect is searched. In the iterative comparison process, a concurrent processor is used, and in the parallel calculation process, the iterative comparison process is developed in a plurality of neighborhood transformation modes.

The control algorithm for optimizing the search preferably adopts a tabu search algorithm, models A1, A2, A3, B1 and B2 all adopt the same method, and pseudo codes are as follows:

1, obtaining an initial solution;

2:repeat

performing neighborhood transformation on the current solution to obtain a neighborhood set;

4

5, if neighborhood solution is superior to the current optimal solution and no tabu list then is listed

6, updating the current solution, the current optimal solution and the tabu list;

7

8

9, using a differential search strategy for the neighborhood solution;

10:end if

finding the excellent solution in the neighborhood set;

12:end if

13:end for

14, if the current optimal solution in the above process is not updated, and the extreme optimal solution in the current neighborhood set is not listed in the tabu list then

Replacing the current solution with the excellent solution in the current neighborhood set and adding the solution into a tabu list;

16:end if

updating algorithm parameters and a tabu list;

18

The route display module is used for displaying the solid waste transportation route according to the decision result of the decision module, preferably presenting the route to a user in a GIS (geographic information system) drawing mode, and scheduling the solid waste transportation by the user according to the result of the route display module;

specifically, the route display module sends the access sequence of each transport node to a vehicle-mounted navigation terminal in a transport vehicle, and the vehicle-mounted navigation terminal serves as an auxiliary transport node for distributed calculation, calculates and displays a transport route and performs navigation; since the cooperative controller in the decision calculation module takes the handling of uncertain conditions such as road congestion into account, the route display module also has the stability of handling uncertain conditions such as road congestion; when serious road congestion occurs or traffic accidents of transport vehicles occur, and the like, the vehicle-mounted navigation terminal is used as an auxiliary transport node for distributed computation, other feasible transport routes are computed and sent to the task acquisition module, and the system is triggered to make decision again for computation.

It will be understood by those skilled in the art that the foregoing is only an exemplary embodiment of the present invention, and is not intended to limit the invention to the particular forms disclosed, since various modifications, substitutions and improvements within the spirit and scope of the invention are possible and within the scope of the appended claims.

Claims

1. A multi-stage collaborative decision-making system for classified transportation of solid wastes based on distributed computation is characterized by comprising a task obtaining module, a model selecting module, a decision-making computation module and a route display module;

the task acquisition module is used for identifying the solid waste transportation mode of the area to be decided, acquiring the solid waste transportation system parameters of the area to be decided and submitting the parameters to the model selection module; the solid waste transport mode comprising a primary solid waste transport mode and a secondary solid waste transport mode; the primary mode of solid waste transport, suitable for transport between transport nodes of level 0 to level 1, comprises: mixed transportation, classified combined transportation, and classified individual transportation; the secondary solid waste transportation mode is suitable for transportation between the 1 st level and above transportation nodes, and comprises: towed transport, and modified towed transport; the solid waste transport system parameters comprising: location and open time of the yard; the number of each level of transportation nodes, the position of each transportation node, the production capacity or the transfer capacity and/or the processing capacity, and the opening time; the number of vehicles, the types of the vehicles, the number of containers, the corresponding loading capacity and the unit mileage transportation cost are used among the transportation nodes; manual configuration of transportation nodes and vehicles at all levels;

the model selection module is used for decomposing a region to be decided into decision units according to the solid waste transportation mode obtained by the task obtaining module, selecting a decision model for each decision unit and submitting the decision model to the decision calculation module; for a decision unit from a 0-level node to A1-level node obtained by decomposing a region to be decided by the model selection module, for a decision unit from the 0-level transportation node to the 1-level transportation node, selecting an A1 model when a mixed transportation mode is used, selecting an A2 model when a classified joint transportation mode is used, and selecting an A3 model when classified independent transportation is used; for the transportation optimization problem from the 1 st-level transportation node to the 2 nd-level transportation node and for the transportation optimization problem from the 2 nd-level transportation node to the 3 rd-level transportation node, selecting the B1 model when using the general drag type transportation mode and selecting the B2 model when using the improved drag type transportation mode;

the A1 model, the A2 model and the A3 model are as follows: the objective function is that the total transportation cost is minimum, and the transportation cost consists of variable cost and fixed cost; the variable cost is linearly related to the total driving range; fixed fee and vehicle usage number K ₁ Linear correlation; the constraint conditions include: degree constraint, vehicle load capacity constraint, driving time constraint, supply and demand balance constraint, time window constraint, sub-circuit elimination constraint and variable value constraint;

wherein, the A1 model vehicle loading capacity constraint function:

in the above formula, m _k Indicating the load capacity, y, of the vehicle k _ik Representing the loading of the vehicle k at the 0 th level transportation node i epsilon C; x is the number of _ijk A non-negative integer variable representing the vehicle K ∈ {1,2 ₁ The number of times the arc (i, j) is traversed;

supply-demand balance constraint function:

in the above formula, y _ik Represents the loading capacity of the vehicle k at the 0 th level transportation node i epsilon C, d _i Represents the transportation demand of the 0 th level transportation node i ∈ C, p _l Representing the service capability of the level 1 transport node l belonging to W;

wherein, the A2 model vehicle loading capacity constraint function:

in the above formula, y _ik And f _ik Respectively representing the quantity of the 2 types of solid wastes loaded by the vehicle k at a 0-th level transportation node i epsilon C, m _k Represents the maximum load capacity of the vehicle k;

supply-demand balance constraint function:

the sub-itinerary elimination constraint and the variable value constraint are classified combined transportation forms;

the target function of the A3 model is the sum of the target functions when the 2 types of solid wastes are respectively modeled according to the A1 model;

the B1 model in the model selection module is as follows:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is the number of _ijk A non-negative integer variable representing the vehicle K ∈ {1,2 ₂ The number of times the arc (i, j) is traversed;

the B2 model in the model selection module is as follows:

in the above formula, D is a yard point set, W is a set of nth-level transportation nodes (n is more than or equal to 1), F is a set of n + 1-level transportation nodes (n is more than or equal to 1), and K is ₂ Number of vehicles used; x is the number of _ijk Is a nonnegative integer variable representing that the vehicle K is in an E {1,2 ₂ The number of times arc (i, j) is experienced;

the decision calculation module is used for performing distributed calculation-based collaborative optimization on the area to be decided according to the solid waste transportation system parameters acquired by the task acquisition module and the decision model selected by the model selection module, so as to acquire a decision result;

the decision calculation module comprises a parameter mapping submodule and at least one model calculation submodule;

the parameter mapping submodule is used for mapping various parameters among a train yard, an nth-level transportation node and an n +1 th transportation node in a decision unit decomposed by the model selection module to the model selection system according to the parameters of the solid waste transportation system obtained by the task obtaining module, taking the parameters as model parameters and submitting the parameters to the model calculation submodule;

the model calculation submodule is used for calculating a transportation route optimization decision model by using a heuristic algorithm according to the decision model selected by the model selection module and the model parameters obtained by mapping of the parameter mapping submodule;

and the route display module is used for displaying the solid waste transportation route according to the decision result of the decision calculation module.

2. The distributed computing-based multi-stage collaborative decision making system for solid waste sorting transportation according to claim 1, wherein the model computation submodule comprises a sequential processor, at least one concurrent processor, a collaborative controller;

said sequential processor, preferably employing an operator comprising a CPU, for calculating an initial solution to said solid waste transport system, i.e. a set of decision variables in an optimization model; controlling the at least one concurrent processor to perform optimization search according to the initial solution; when the optimization search operation is completed or the jump-out condition is met, obtaining a final solution of the solid waste transportation system according to a result returned by the concurrent processor;

the cooperative controller is used for cooperatively controlling the transportation process from the nth-level transportation node to the (n + 1) th-level transportation node and the transportation process from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node, wherein the two transportation processes are intersected at the (n + 1) th-level transportation node; the cooperative controller can process the solid waste transportation system parameters acquired by the task acquisition module, and increase the service time or the residence time of the vehicle at the transportation node so as to enhance the stability of the decision calculation module for coping with uncertain conditions such as road congestion; the cooperative controller is controlled according to the following method: and using a sequential processor and a concurrent processor, firstly, initially determining a time window from the (n + 1) th-level transportation node to the (n + 2) th-level transportation node in the transportation process, secondly, further adjusting and optimizing the time window of the two-level transportation process by simultaneously solving the two-level transportation process for multiple times, and finally obtaining a set of optimal solution of the two-level transportation process in cooperation.

3. The distributed computing-based multi-stage collaborative decision making system for solid waste classification transportation according to claim 1, wherein the decision computation module obtains an initial solution as follows: adding a feasible and current optimal transport node in the current path for the A1 model and the A2 model; and adding a feasible unit path to the B1 model, the B2 model and the A3 model in the current path in sequence, wherein the unit path refers to a process that a vehicle starts from a certain nth-level transportation node after being loaded, unloads to a certain (n + 1) th-level transportation node adjacent to the vehicle and then to the certain nth-level transportation node, and n is more than or equal to 1.

4. The distributed computing-based solid waste classification transportation multi-stage collaborative decision making system according to claim 1, wherein the decision computation module obtains a final solution as follows: starting from the initial solution, performing neighborhood transformation on the current solution for iterative comparison: if the solution after the neighborhood transformation is better than the current solution, replacing the current solution; if the solution after the neighborhood transformation meets the requirement of iterative convergence, the solution is used as the obtained optimal solution; the neighborhood transformation refers to operations such as breaking, crossing and recombining the path sequences of different vehicles in a group of vehicle paths, so that a new group of vehicle paths are obtained.

5. The distributed computing-based solid waste sortation transportation multi-level collaborative decision system of claim 1, further comprising a route display module; and the route display module is used for displaying the solid waste transportation route according to the decision result of the decision calculation module, preferably presenting the solid waste transportation route to a user in a GIS (geographic information system) drawing mode, and scheduling the solid waste transportation by the user according to the result of the route display module.