Network Reliability

Network robustness and network reliability are important issues in the design of Internet Service Providers’ topologies. In this paper, we examine the structural characteristics of network topologies that affect robustness and reliability. We examine the interplay between the structural characteristics of network topologies and the resource capacity over-provisioning strategies when the network breakdowns subject to practical constraints (router technology) and economic considerations (link costs). We study the robustness of the Internet connectivity under node intentional harmful attack using two attacks strategies: static degree-based and static load-based. We find that the robustness of network topologies is affected by the variation of their structural characteristics. In our proposed approach, we show that highly-heterogeneous topologies have less robustness compared with lightly-heterogeneous topologies. The observations from the robustness study provide us useful insights for proposing multiple efficient preventive resource capacity over-provisioning strategies for mitigation of intentional attacks. The proposed strategies utilize the structural properties by calculating the excess traffic in case of single global cascading failure for each node and measure its influence on the other nodes as well as locally. The results show that our proposed strategies can significantly enhance the robustness and increase the resilience of network topology. We also show that highly-heterogeneous topologies have high resilience compared with lightly-heterogeneous topologies. By using real data from the Sprint network at the router level, we provide further empirical evidence in support of the proposed approach.

This paper describes the development of a method for strategic assessment of vulnerability in road networks. Our vulnerability analysis considers the socio-economic impacts of network degradation, on the basis of changes in the levels of accessibility provided by the degraded network. A number of standard indices of accessibility are considered, including generalised travel cost, the Hansen integral accessibility index, and the ARIA index used in Australia to assess levels of remoteness in rural areas. Simple case studies of a national strategic road network and a regional road network are used to illustrate both the general approach and specific implementations.

Abstract: The network topological design problem (NTDP) is a combinatorial optimization problem and it is NP-hard so that the solution time using classical enumeration-based methods grows exponentially with the network size. The problem considered in this study is to find the best layout of links between a given set of nodes, such that at least 2-connectivity is not violated, desired network reliability is satisfied and total cost is minimized. We discuss the development of a Tabu Search (TS) method for NTDP considering a short term memory structure. An experimental design analysis over networks of various sizes and configurations is made in order to choose the best values of the algorithm parameters. We establish the effectiveness of our approach by comparing its performance with regard to the number of solutions examined in the search space and the solution quality relative to Genetic algorithm (GA) heuristic and optimization method. The results show that, the TS approach is compu...

The paper presents a multi-objective planning approach for electrical distribution systems incorporating shunt capacitor banks. In this planning, the number of feeders, feeder routes, and number, locations, and rating of shunt capacitor banks for a distribution system are determined using a multi-objective optimization approach. The objectives considered for this optimization are: (i) total investment cost minimization, (ii) network reliability maximization, (iii) minimization of network power loss, and (iv) minimization of node voltage deviation. The last three objectives are aggregated into a single objective, named as network performance metric. This objective and the first objective are simultaneously optimized to obtain a set of non-dominated solutions. The solution algorithm is Multi-objective particle swarm optimization (MOPSO). Three different MOPSO variants are used and their performances are compared. The result of this planning approach is compared with that of the planning without capacitor banks. The proposed planning approach is validated on a 54-node distribution system planning problem.

We study the problem of transmission-side diversity and routing in a static wireless network. It is assumed that each node in the network is equipped with a single omnidirectional antenna and that multiple nodes are allowed to coordinate their transmissions in order to obtain energy savings. We derive analytical results for achievable energy savings for both line and grid network

We show how binary decision diagrams (BDDs) can be used to solve and obtain postoptimality analysis for linear and nonlinear integer programming problems with binary or general integer variables. The constraint set corresponds to a unique reduced BDD that represents all feasible or near-optimal solutions, and in which optimal solutions correspond to certain shortest paths. The BDD can be queried in real time for in-depth postoptimality reasoning. The approach is equally effective for linear and nonlinear problems. There are currently no other methods for obtaining such an analysis, short of repeatedly re-solving the problem. We illustrate the analysis on capital budgeting and network reliability problems.