CN100492224C - Power management method for electronic equipment - Google Patents

Abstract

A power management method of electronic appliance includes using Bayes estimation network to predict user's operation state, using operation system obtain percentage number of cell power, inputting operation state probability number and cell power percentage number into fuzzy controller for calculating out wait close time of each electronic appliance, accumulating user's continuous unresponse time and cutting power off for saving energy if said unresponse time is over the waiting close time calculated out by fuzzy controller.

Description

Power management method for electronic equipment

technical field

The present invention relates to a power management method of electronic equipment, more specifically to a power management method of electronic equipment for energy saving.

Background technique

Energy-saving management of power supply is an effective means for energy-saving control of electronic equipment. Existing electronic equipment power management only formulates power management schemes based on user-set waiting time values for each electrical equipment, so as to achieve the purpose of energy saving. In actual use, it is up to the user to set according to the power mode preset options, for example, the user selects a laptop computer in the power mode, waits for 15 minutes to enter the power saving state, and so on. The energy-saving effect of the existing electronic equipment power management method is not obvious, and there are the following deficiencies: 1. The existing power management method is to formulate a power management plan based on the user's setting of the waiting time value for each electric device, without considering the user The actual situation during the working state of the equipment. 2. The power management method is mechanical and single, and cannot adapt to the changing working state of the user. 3. Users usually do not care about the operation of power management during use, and cannot adjust the power scheme in time according to needs, so that the various power management modes initially set cannot play their role, and the effect of energy saving cannot be achieved. 4. If the time set by the user is longer, then when the frequency of use of the device by the user is low, the idle waiting time of the device is too long, resulting in unnecessary waste of electric energy. 5. If the time set by the user is short, when the user is in a high frequency of equipment use, the equipment will be restarted frequently, which will increase the consumption of electric energy and easily cause damage to the equipment.

Contents of the invention

The technical problem to be solved by the present invention is to provide a new type of electronic equipment power management method, which adopts the dynamic Bayesian network probabilistic reasoning technology and fuzzy control technology in the intelligent technology, and combines the current working state of the user and the battery of the current equipment by the system. Power, predict and automatically adjust the waiting time of the device according to the time range set by the user, and minimize the waste of electric energy caused by the user inadvertently, so as to achieve the purpose of saving energy and extending the working time of electronic devices under battery conditions.

The technical solution adopted in the present invention: a power management method for electronic equipment, comprising the following steps:

The first step is to use the dynamic Bayesian network probabilistic reasoning technology to infer the working status of the user's equipment. The final calculation formula is:

$P\left(W_t\right|r_{1:t})=\mathrm{\αP}\left(r_t\right|W_t)\underset{w_{t-1}}{\mathrm{\Σ}}P\left(W_t\right|w_{t-1})P\left(w_{t-1}\right|r_{1:t-1})$ (Equation 1)

The P(W _t |r _1:r ) is the probability value to be calculated to reflect the working state of the device used by the user at the current moment, the P(r _t |W _t ) is the mathematical calculation model of the sensor, and the P(W _t |w _t-1 ) is the transfer mathematical calculation model, the P(w _t-1 |r _{1: t-1} ) is the probability value of the working state of the device used by the user at the previous moment, and the α is the transfer mathematical calculation model. Calculation model, sensor mathematical calculation model and the results obtained from the calculation of the probability value of the user's working state of the device at the previous moment are converted to a normalization constant in the range of 0-1, and the 1:t represents an integer sequence from 1 to t , including 1 and t, the formula of the S-curve function is:

$f\left(t\right)=\frac{b-a}{1+e^{-\ln \left(\frac{b-a-\mathrm{\ρ}}{\mathrm{\ρ}}\right)*(t-\mathrm{\μ})/\mathrm{\μ}}}+a$ (Equation 2)

The variable of the S-shaped curve function is t, the lower limit of the curve is a, the upper limit of the curve is b, the center value is μ, and the error is ρ. The values of the transfer mathematical calculation model and the sensor mathematical calculation model are adjusted within a certain range to fit the user's use The working status of the equipment;

The second step is to call the underlying function of the operating system of the electronic device to monitor the battery power percentage value of the device;

The third step is to use fuzzy control to combine the working status of the user's equipment obtained in steps 1 and 2 and the battery power percentage value of the equipment to formulate the waiting time for closing the equipment. The fuzzy control is composed of a fuzzy mathematical calculation model and a fuzzy inference engine The mathematical calculation model consists of four parts: the fuzzy rule library and the mathematical calculation model of the defuzzer: the mathematical calculation model of the fuzzer uses a quadrilateral fuzzer to construct a membership function μ(x) to calculate two language variables, that is, the The probability value of the working state and the fuzzy value of the battery power percentage value of the device. In the mathematical calculation model of the fuzzy inference machine, the fuzzy inference machine uses the product inference machine as a calculation method. First, the fuzzy rule base is established, and its rule form is: If S=s1 and B=b1 Then W=w1, the If part is called the antecedent of the rule, and the Then part is called the latter, S, B, W are all fuzzy quantities, and the number of rules in the fuzzy rule base depends on the two The product of the number of language values corresponding to each language variable, and then use the probability value of the user's working state of the device and the fuzzy value of the battery power percentage value of the device to obtain the corresponding waiting time for the device to be turned off according to the fuzzy rule base constructed. The fuzzy value of the fuzzy value, the calculation model of the defuzzifier adopts the calculation method of the center average defuzzification to weight the multiple fuzzy values within the scope of the domain constructed, and obtain the only value of the waiting time for the device to be closed, and the weighted average formula for:

$\mathrm{the\; y}=\frac{\underset{l=1}{\overset{m}{\mathrm{\Σ}}}\stackrel{\‾}{{\mathrm{the\; y}}^l}{w}_l}{\underset{l=1}{\overset{m}{\mathrm{\Σ}}}{w}_l}$ (Equation 3)

In the weighted average formula, y ^l is the central value of the L fuzzy set, and w _t is a fuzzy value corresponding to the equipment waiting time value for closing;

The fourth step is the execution of power management: after the time set by a transfer time interval, repeat the first and second steps and refresh the waiting time of each device saved at the previous moment according to the working status of the device used by the user Data, the data calculated at the current moment is used as the new judgment standard, and at the same time, the time since the user's last response to the device is counted, that is, the non-response time. When the non-response time exceeds the waiting time for the device to be closed, the Then start the energy saving management for the device, turn off the power of the device through the circuit to implement energy saving, until the user responds to the electronic device again, turn on the power of the device through the circuit, and return to the working state of the device used by the user.

Beneficial effects of the present invention: the present invention uses the probabilistic reasoning technology and fuzzy control technology of the dynamic Bayesian network in the intelligent technology to predict and automatically adjust the waiting time of each device in combination with the user's current working status of the device and the battery power of the current device. Closing time. The present invention refers to the current power condition, and when the device is in continuous use, prolongs the waiting time of the device in a timely manner; when the user's working frequency is low, the suspension time is automatically shortened. In this way, it eliminates the phenomenon of wasting power caused by the device waiting for too long to shut down when the frequency of use is low, saves the power of the device, and solves the problem that the device may be started multiple times when the time setting is too short problems and increase the service life of the equipment. The present invention can be widely used in the power management of modern electronic equipment, especially for the battery power management of portable computers, the present invention can greatly improve the energy saving degree and performance of the equipment. Specifically, in the high working state and high power, the present invention always sets the power scheme at a higher waiting and suspending time, and the occasional idle interval will not affect the system to judge that the user is still working, and there will be no high The phenomenon of frequently turning off and turning on electrical equipment achieves better performance; under high working conditions and lower power levels, compared with the previous situation, the waiting and suspending time configured by the present invention is more moderate , this is because the battery power factor has become an important indicator of the present invention's evaluation, so as to shorten the time as much as possible, thereby saving electric energy. Correspondingly, in a low working state and high power, the present invention configures the power supply scheme in a middle-lower time range most of the time, that is, the waiting and suspending time is shorter. Avoid unnecessary waste of electric energy in the idle state, and play the effect of power management; similar to the second case, when the present invention detects a low working state and low battery, the system will naturally shorten the waiting time of the device to the minimum , so that the waste of electric energy is reduced to a minimum to extend the battery life. It can be seen that through the automatic power supply management of electronic equipment in the present invention, the power supply energy saving control is adapted to the user's work changes and the battery power of the equipment, taking into account both equipment energy saving and performance.

Description of drawings

Fig. 1 is the sigmoid function curve diagram corresponding to the numerical variation range of the transfer mathematical calculation model;

Fig. 2 is the S-type function curve diagram corresponding to the numerical variation range of the sensor mathematical calculation model;

Fig. 3 is the line graph (membership degree function graph) of each language value range corresponding to the working status probability value of the user's equipment in the described fuzzer;

Fig. 4 is the line diagram (membership function diagram) of each language value range corresponding to the percentage value of battery power in the fuzzer;

Fig. 5 is a line diagram (membership function diagram) corresponding to each language value range corresponding to the device waiting for closing time (short) in the defuzzifier;

Fig. 6 is a line diagram (membership function diagram) corresponding to each language value range corresponding to the device waiting for closing time (shorter) in the defuzzifier;

Fig. 7 is the broken line diagram (membership function diagram) corresponding to each language value range corresponding to the device waiting shutdown time (general) in the defuzzifier;

Fig. 8 is a broken line diagram (membership degree function diagram) corresponding to each language value range corresponding to the device waiting for closing time (longer) in the defuzzifier;

Fig. 9 is a broken line diagram (membership degree function diagram) corresponding to each language value range corresponding to the device waiting for closing time (long) in the defuzzifier;

Fig. 10 is a flowchart of the present invention;

Fig. 11 is a structural diagram of dynamic Bayesian network probabilistic reasoning;

Figure 12 Fuzzy control process diagram.

Fig. 13 is a work interface diagram 1 based on the power management software of the present invention;

Fig. 14 is the working interface drawing 2 of the power management software based on the present invention.

Detailed ways

Below by accompanying drawing the present invention is described in further detail:

The first step is to use the probabilistic reasoning technology of the dynamic Bayesian network to infer the working status of the user's equipment.

(1) Estimation method of user usage status

The probabilistic reasoning of the dynamic Bayesian network is to use the user's activities in the past stage (that is, the user's response to the electronic device) to dynamically predict the possibility of the user's current working state, and express it with a probability value, which is calculated by the transfer mathematics Model, sensor mathematical calculation model and transfer time interval value constitute the whole calculation process.

Calculation formula:

$P\left(W_t\right|r_{1:t})=\mathrm{\αP}\left(r_t\right|W_t)\underset{w_{t-1}}{\mathrm{\Σ}}P\left(W_t\right|w_{t-1})P\left(w_{t-1}\right|r_{1:t-1})$ (Formula 1)

In (Formula 1), P(W _t |r _1:t ) is the probability value to be calculated to reflect the working state of the device used by the user at the current moment; P(r _t |W _t ) is the mathematical calculation model of the sensor; P(W _t |w _t-1 ) is the transition mathematical calculation model; P(w _t-1 |r _{1: t-1} ) is the probability value of the working state of the user's equipment at the previous moment. α is a normalization constant, which converts the results obtained from the calculation of the transfer mathematical calculation model, the sensor mathematical calculation model, and the probability value of the working state of the device used by the user at the previous moment to a probability value in the range of 0-1; 1: t means from An integer sequence from 1 to t, including 1 and t, that is, all sequences from the first calculation to the present. (Formula 1) is a recursive calculation process composed of the transfer mathematical calculation model, the sensor mathematical calculation model and the conditional probability distribution of the calculation results at the previous moment. The system calculates the user's usage status at the same transfer time interval, and This calculation result is used as a calculation factor for the next calculation, so that (Formula 1) can be used to form a real-time probabilistic reasoning process that detects and calculates the user's usage status.

(2) Determination of transfer mathematical calculation model, sensor mathematical calculation model and transfer time interval value.

The transfer mathematical calculation model and the sensor mathematical calculation model in (Formula 1) are two key data items, which will directly affect the final reasoning result. Therefore, constructing these two models according to the actual situation of the environment in which the model is located will help the probabilistic reasoning to be more in line with the actual situation and achieve a prediction effect that can reflect the real usage state. In the present invention, adopt a kind of S-type curve mathematical model, construct a S-type function curve about time t as a soft threshold value through deformation according to the requirement of calculation, see (formula 2), make transfer mathematical calculation model and The value of the sensor mathematical calculation model can be adjusted within a certain range to fit the user's situation when using the device.

$f\left(t\right)=\frac{b-a}{1+e^{-\ln \left(\frac{b-a-\mathrm{\ρ}}{\mathrm{\ρ}}\right)*(t-\mathrm{\μ})/\mathrm{\μ}}}+a$ (Formula 2)

Sigmoid function: variable t; curve lower limit a; curve upper limit b; central value μ; error ρ

1). Transfer mathematical calculation model:

The transfer mathematical calculation model is a law that describes how the working state of the user's equipment evolves over time, that is, when the user is working in the previous state, there is a possibility of working in the current state. If the user responded to the device in the previous state, there is a higher probability that the user will continue to respond to indicate work at the current moment, and vice versa. As the working time evolves, the system will gradually think that the possibility of the user continuing to maintain this state will decrease when the user uses it for a long time, that is, the higher probability value will decrease. On the contrary, in the case of not using it for a long time, the probability of using the device in the future will gradually increase, that is, the lower probability value will increase, see Table 1.

Table 1. Numerical change table of the transfer mathematical calculation model

W_t-1 P(W_t=t) P(W_t=f) t 0.9-0.7 0.1-0.3 f 0.1-0.3 0.9-0.7

The whole process can be represented by an S-shaped curve, as shown in Figure 1. After each execution of turning off the power of the device, the value of the transfer model will be restored to indicate the end of the user's work at this stage and enter a period of rest. With such a functional factor about the total time of use, it is more accurate to predict the user's use status.

2) Sensor mathematical calculation model:

The sensor mathematical calculation model describes how the user's response to the device is affected by the user's working state of the device, that is, the user's response to the device can be regarded as a performance of the user's working state of the device. When the system detects that the user's continuous response state (the user has been responding to the device or the user has not responded to the device) changes to another state, the system's confidence in the value of the sensor model obtained by the change is low. With the continuation of this state, that is, the same state continues to appear several times, the system will gradually increase the reliability, thinking that the value obtained by the sensor model is relatively correct, so the corresponding probability will gradually increase, see Table 2 .

Table 2. Numerical changes of the sensor mathematical calculation model

W_t P(R_t=t) P(R_t=f) t 0.6-0.9 0.4-0.1 f 0.4-0.1 0.6-0.9

The whole process also uses an S-shaped curve to represent the numerical change, as shown in Figure 2. Doing so can better conform to the real usage situation of the user, eliminate the interference caused by some accidental factors, and improve the accuracy of predicting the working status of the device used by the user.

3) Transfer time interval value

The transfer time interval value means the time interval between the system's two judgments on the working status of the user's equipment. Due to the different characteristics of each electronic equipment, the value can also be different, generally within a short period of time.

The second step is to call the underlying function of the operating system of the electronic device to monitor the battery power percentage value of the device;

The third step is to use fuzzy control technology to combine the working status of the user's equipment with the battery power information of the equipment to formulate the waiting time for the equipment to be turned off. Fuzzy control is a rule-based control system used to deal with the nonlinear relationship between user usage status, device battery power and device waiting time to turn off. Fuzzy control is composed of four parts: fuzzer mathematical calculation model, fuzzy inference machine mathematical calculation model, fuzzy rule library and defuzzifier mathematical calculation model. Since the system calculates and determines the waiting time value of the device by combining the user's usage status and the device's battery power, fuzzy calculations are required for the probability value of the user's usage status and the percentage of the device's battery power to obtain the corresponding device waiting time for shutdown.

1) Calculation of the fuzzer:

The fuzzer system uses a quadrilateral fuzzer to construct the membership function μ(x). The advantage is that the quadrilateral fuzzer can overcome the noise contained in the input variables, and it can also simplify the calculation process. See Figure 3 and Figure 4. The user uses the state probability The fuzzy linguistic variables corresponding to the values have 7 linguistic values in total, and these 7 linguistic values constitute a set of quadrilateral membership functions. The numerical range of fuzzy linguistic variables and linguistic values is called the domain of discourse. Similarly, the battery percentage value also corresponds to 5 language values. The entire calculation process of the fuzzer is to use the probability value of the user's use state and the battery power percentage value of the device to correspond to the degree of membership of each language value in the constructed language variable, that is, according to the membership function of the input variable, find the exact The input value of the input value corresponds to the membership degree of each language value of the input variable, and the fuzzification process in the entire fuzzy calculation is completed. is obtained by the fuzzer.

2) Calculation of fuzzy inference engine:

First of all, according to the actual situation, rely on the experience and knowledge of experts to establish a fuzzy rule base. It is the core of fuzzy calculation, and the form of a rule is: If S=s1 and B=b1 Then W=w1, where the If part is called the antecedent of the rule, and the Then part is called the latter, and S, B, and W are all Amount of blur. The number of rules in the rule base depends on the number of language values contained in each input variable, that is, the product of the number of language values in the language variables corresponding to the user's use state probability value and the device battery power percentage value, see Table 3.

Table 3. Fuzzy rule base

Then, the probability value of the user's use state obtained in the previous step and the fuzzy value of the battery power percentage value of the device are used to obtain the corresponding fuzzy value of the waiting time for the device to be turned off according to the constructed fuzzy rule base. The fuzzy inference machine uses the product inference machine as the calculation method, that is, the fuzzy values of the two antecedents of each rule are multiplied, and the resulting value is used as the strength of each rule in the rule base, and then the rules with the same consequent Take the maximum value of its intensity as the fuzzy value of this kind of consequent. In this way, according to the difference in the number of successors in the effective rules, that is, the difference in the language value of the output language variable, the fuzzy inference engine will output multiple fuzzy values about the waiting time for the device to be turned off, and complete the work of the user using the device reflected by the fuzzy rule base The corresponding relationship between the state probability value and the battery power value of the device and the waiting time for the device to be turned off.

3) Calculation of the defuzzifier:

The defuzzification calculation adopts the calculation method of center average defuzzification. Since there are multiple fuzzy values in the fuzzy output value about the waiting time of the equipment in the fuzzy reasoning machine, it is necessary to use the center average to defuzzify, and to weight the average of multiple fuzzy values within the constructed universe to obtain Its only one device wait time value, see Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, the range of each language value corresponding to the device wait time (membership function), a total of 5 kinds of device wait time Range, Short, Shorter, Normal, Longer, Long.

The weighted average formula is:

$\mathrm{the\; y}=\frac{\underset{l=1}{\overset{m}{\mathrm{\Σ}}}\stackrel{\‾}{{\mathrm{the\; y}}^l}{w}_l}{\underset{l=1}{\overset{m}{\mathrm{\Σ}}}{w}_l}$ (Formula 3)

In (Formula 3), y ^l is the central value of the L fuzzy set, and w _t is a fuzzy value corresponding to the waiting time value of the device. After calculating the probability value based on the user's usage status and the battery power percentage value of the device through the defuzzifier to calculate the waiting time value of the device, the entire fuzzy calculation process is completed, that is, the power management scheme of the electronic device can be automatically determined. work out.

The fourth step is the execution of power management: the system repeats the first, second, and third steps of predicting the working status of the user's equipment and waiting for the set waiting time to be turned off every time the system passes through the time set by the transfer time interval. The process of formulating the waiting time for the device to be closed within the scope, refresh the data of the waiting time for each device saved at the previous moment according to the user's response, and use the data calculated at the current time as the new judgment standard. When the user responds to the device within the time set by the transfer time interval, the waiting time of the device stored in the system will be continuously updated to conform to the power management scheme corresponding to the current state of the user. If the user does not respond to the device within this time, the device waiting time saved in the system will not be replaced, and the existing data is the power management scheme corresponding to the state of the user's last response, which is used to judge The basis for whether the user's continuous non-response time exceeds the device's waiting time for shutdown. At the same time, the system counts the time since the last time the user responded to the device, that is, the unresponsive time. When the continuous non-response time exceeds the waiting shutdown time of the device obtained by the system, the energy saving management for the device is started, and the power supply of the device is turned off through the circuit to implement energy saving. Until the user responds to the electronic device again, wakes up the device, turns on the power of the device through the circuit, and resumes the working state.

Example:

The invention has been applied to power management software for mobile computers and desktop computers. Due to the adoption of dynamic Bayesian network probabilistic inference technology and fuzzy control technology, the original single and mechanical management mode of power management has become a more intelligent and humanized flexible management mode, especially its energy-saving degree and performance Both are improved compared to traditional power management, so that in the current situation of high battery power requirements for mobile electronic devices, through power management methods, the power saving control is adapted to the user's needs without losing the performance of the device. Work changes and the battery power of the device can minimize the waste of power caused by the user inadvertently, thereby improving the power supply efficiency of the battery and extending the working time of the device. Aiming at the specific use of mobile computers and desktop computers, the invention is applied to produce intelligent power management software, and energy-saving control of computer equipment is carried out through the power management software. Start the power management software, and within the transfer time interval (the time value set in this example is 1 minute), the software will obtain the information of whether the user has a response device, that is, the user moves the mouse or taps the keyboard and other input devices, and then the software according to The mathematical calculation model of the sensor selects a calculation value that can represent this response situation as a calculation factor for the entire prediction, that is, according to the duration of the user's equipment response, the corresponding probability value is obtained by using the S-curve function, see Table 2. Figure 2. At the same time, as the user's use of the device increases, the software also selects a numerical value that can represent the change of the current user's usage in the transfer mathematical calculation model as another calculation factor for the entire prediction, that is, according to the continuous working state of the user's use of the device Time, using the S-shaped curve function to obtain the corresponding probability value, see Table 1 and Figure 1. Finally, the software multiplies the determined transfer model value and sensor model value with the probability reasoning value at the previous moment to obtain the probability value for quantifying and predicting the working state of the user's device at the current moment, as shown in Figure 11. Figure 11 is a structural diagram of the dynamic Bayesian network probabilistic inference, 1. The process of changing the working status of the user's equipment with the evolution of time. 2. Transfer the mathematical calculation model to describe how the working status of the user's equipment evolves over time. 3. The user's response to the input device. 4. A mathematical computational model of the sensor that describes how the user's response to the device is affected by the user's work situation. After the software obtains the probability value of the working state of the user's device at the current moment through the probabilistic reasoning of the dynamic Bayesian network, the software monitors the current battery power of the mobile computer and obtains the corresponding battery power percentage value. The probability value of the working state of the user using the device and the percentage value of the computer battery power are input into the fuzzy calculation model, as two precise input quantities, through the fuzzy device calculation model (the set membership function is shown in Fig. 3, Fig. 4), the fuzzy inference computer calculation model (see Table 3 for the set fuzzy rule base) and the defuzzifier calculation model (see Figure 5, Figure 6, and Figure 7 for the set membership function and the waiting time range of five kinds of equipment , Fig. 8, Fig. 9), and finally obtain the waiting time value of a certain device of the computer, such as a monitor, a hard disk, etc. In order to improve the applicability of the software and enhance the user's choice, the software presets 5 different equipment waiting time ranges for the user to choose from various electrical equipment, that is, the software waits to be shut down according to the equipment selected by the user during the calculation process. For the time range, input the fuzzy value obtained by the fuzzy inference engine into the mathematical calculation model of the defuzzifier, obtain the waiting time for the equipment within the range selected by the user, and save the obtained data, as shown in Figure 12. 12 is a diagram of the fuzzy control process, 1. The probability value of the working state of the device used by the user and the value of the battery power of the device obtained by the dynamic Bayesian network probabilistic reasoning and detection of the battery power of the device. 2. The fuzzer, which fuzzifies the input precise quantity through the quadrilateral function, and outputs the corresponding fuzzy value as the data calculated by the inference engine. 3. Fuzzy inference engine, using product inference engine to calculate fuzzy rules. 4. The fuzzy rule base reflects the waiting time for the device to be turned off corresponding to the probability value of the working state of the device used by the user and the value of the battery power of the device. 5. The defuzzifier, which defuzzifies according to the quadrilateral function, and calculates the waiting time for the device to be turned off. 6. According to the waiting time range set by the user, determine the waiting time of each device and configure the power management scheme. The software counts the time since the last time the user responded to the device, that is, the unresponsive time. When the continuous non-response time exceeds the waiting time for the device to be shut down obtained by the fuzzy control calculation model, the software starts the energy-saving management of the device, and the power of the device is turned off through the circuit to implement energy-saving. Until the user responds to the electronic device again and wakes up the device, the software turns on the power of the device through the circuit and resumes the working state, as shown in Figure 10. Figure 10 is a structural diagram of power management software based on intelligent technology, 1. Use the probabilistic reasoning of dynamic Bayesian network in intelligent technology to predict the user's working status. 2. Obtain information about the battery level of the computer through the operating system. 3. Input the obtained working state probability value of the equipment used by the user and the electric quantity value of the computer battery into the fuzzy controller, and calculate the waiting shutdown time of each electric equipment. 4. The software accumulates the user's continuous non-response time. When the continuous non-response time exceeds the waiting time for the device to be turned off obtained by the fuzzy controller, the software turns off the power of the device to implement energy saving for the device. As a general power management method, the present invention has lower requirements on system hardware. Since only relying on the user's response information to the device and the battery power of the device, dynamic state prediction and power management can be realized, so only a few key parameters in the model, transfer time interval, sensor model, transfer model , fuzzy quantity and fuzzy rule library can be applied in other mobile electronic devices such as handheld computers and smart phones after proper modification to realize the promotion and application of this power management technology, which has high commercial development and utilization value. FIG. 13 is the working interface diagram 1 of the power management software based on the present invention, and FIG. 14 is the working interface diagram 2 of the power management software based on the present invention.

The above content is only a basic description of the concept of the present invention, and any equivalent transformation made according to the technical solution of the present invention shall fall within the scope of protection of the present invention.

Claims (1)

1. the method for managing power supply of an electronic equipment comprises the following steps:

The first step, utilization dynamic bayesian network probability inference technology infers that the user uses the duty of equipment, total formula is:

$P\left(W_t\right|r_{1:t})=\mathrm{\αP}\left(r_t\right|W_t)\underset{w_{t-1}}{\mathrm{\Σ}}P\left(W_t\right|w_{t-1})P\left(w_{t-1}\right|r_{1:t-1})$ (formula 1)

Described P (W _t| r _1:t) be that the reflection current time user that will calculate uses the probability numbers of the duty of equipment, described P (r _t| W _t) be sensor mathematics computing model, described P (W _t| W _T-1) for shifting mathematics computing model, described P (w _T-1-r _1:t-1) result of the duty probability numbers calculating gained of equipment converts the normaliztion constant in the 0-1 scope, described 1:t represents from 1 to t integer sequence for the user that will shift mathematics computing model, sensor mathematics computing model and previous moment uses for the user of previous moment uses the duty probability numbers of equipment, described α, comprise 1 and t, S type curvilinear function formula is:

$f\left(t\right)=\frac{b-a}{1+e^{-\ln \left(\frac{b-a-\mathrm{\ρ}}{\mathrm{\ρ}}\right)*(t-\mathrm{\μ})/\mathrm{\μ}}}+a$ (formula 2)

The variable of S type curvilinear function be limited under t, the curve be limited to b on a, the curve, central value is that μ, error are ρ, shift the numerical value of mathematics computing model and sensor mathematics computing model and in certain scope, regulate the duty of using equipment with the match user;

In second step, call electronic equipment operating system bottom function, the battery electric quantity percentages of monitoring equipment;

The 3rd step, the utilization fuzzy control is with step 1, two users that obtain use the battery electric quantity percentages of the duty of equipment and equipment to combine wait shut-in time of formulation equipment, described fuzzy control is by the fuzzy device mathematics computing model, the indistinct logic computer mathematics computing model, fuzzy rule base is conciliate fuzzy device mathematics computing model four parts and constituted: described fuzzy device mathematics computing model adopts the quadrilateral fuzzy device to construct membership function μ (x) and calculates two linguistic variables, it is the fuzzy value of the battery electric quantity percentages of the probability numbers of user's duty of using equipment and equipment, indistinct logic computer uses the product inference machine as account form in the described indistinct logic computer mathematics computing model, at first set up described fuzzy rule base, its rule format is: If S=s1 and B=b1Then W=w1, If partly is called the former piece of rule, Then partly is called consequent, S, B, W is a fuzzy quantity, regular quantity in the fuzzy rule base depends on that described two linguistic variables distinguish the product of corresponding language value quantity, then the user is used the fuzzy value of the battery electric quantity percentages of the probability numbers of duty of equipment and equipment to obtain the fuzzy value of waiting for the shut-in time accordingly about equipment according to the fuzzy rule base correspondence of structure, the computing method that described ambiguity solution device computation model adopts the average ambiguity solution in center in the domain scope of being constructed with a plurality of fuzzy value weighted means, obtain its unique equipment and wait for the shut-in time value, the weighted mean formula is:

$y=\frac{\underset{l=1}{\overset{M}{\mathrm{\Σ}}}\stackrel{\‾}{y^l}{w}_l}{\underset{l=1}{\overset{M}{\mathrm{\Σ}}}{w}_l}$ (formula 3)

In the weighted mean formula, y ^lBe the central value of L fuzzy set, w _lWait for the shut-in time for equipment and be worth a pairing fuzzy value;

The 4th step, the execution of power management: every through a time that sets at interval transfer time, just repeat the first step and second step and use the duty of equipment to refresh each equipment wait shut-in time data that previous moment is preserved according to the user, the data of current time being calculated gained are as new criterion, meanwhile add up the last response apparatus of the user time so far, i.e. response time not, when response time has not surpassed equipment that system tries to achieve and has waited for the shut-in time, then start administration of energy conservation for this equipment, closing the power supply of this equipment by circuit implements energy-conservation, respond electronic equipment once more until the user, open device power supply (DPS) by circuit, return to the duty that the user uses equipment.

Images