Decision-making method and device based on multi-objective optimization and terminal equipment

Abstract

The invention is suitable for the technical field of data processing, and provides a decision-making method and device based on multi-objective optimization and terminal device.The method comprises the steps that decision-making parameters of a to-be-solved multi-objective optimization problem are obtained, and the decision-making parameters at least comprise multiple objective functions influencing decision making and the priority order of the multiple objective functions; randomly generating an initial decision-making population containing a plurality of decision-making schemes based on the plurality of objective functions and the priority order of the plurality of objective functions; according to the priority feature value corresponding to at least one decision scheme in the initial decision population, constructing an iterative decision population; if it is detected that the number of construction times of the iterative decision population reaches the number of algorithm iterations, the iterative decision population is determined as a target decision population, and the target decision population is used for solving the to-be-solved multi-target optimization problem of the decision parameters. According to the method, the optimal decision-making scheme meeting the priority requirement is provided for the user, and the performance of the decision-making method is improved.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular, relates to a decision-making method, device and terminal equipment based on multi-objective optimization. Background Art

[0002] Multi-objective optimization problems with priorities are widely present in many real-world problems such as vehicle design, post-disaster management, and industrial scheduling. Among them, multi-objective optimization problems with priorities refer to optimization problems in which there are multiple conflicting or influencing objective functions, and multiple objective functions participate in decision-making at different priority levels.

[0003] However, when solving such multi-objective optimization problems with priorities, traditional decision-making methods cannot efficiently provide the best decision-making solutions that meet user priority requirements, which reduces the performance of decision-making methods. Summary of the invention

[0004] The embodiments of the present application provide a decision-making method, apparatus and terminal device based on multi-objective optimization, which can provide an optimal decision-making solution that meets the user's priority requirements and improve the performance of the decision-making method.

[0005] In a first aspect, an embodiment of the present application provides a decision-making method based on multi-objective optimization, including:

[0006] Obtaining decision parameters of the multi-objective optimization problem to be solved, the decision parameters at least including: multiple objective functions affecting the decision, and the priority order of the multiple objective functions;

[0007] Based on multiple objective functions and the priority order of the multiple objective functions, an initial decision population including multiple decision plans is randomly generated;

[0008] According to the priority eigenvalue corresponding to at least one decision scheme in the initial decision population, an iterative decision population is constructed, wherein the priority eigenvalue is the eigenvalue obtained by performing non-dominated sorting on the objective function corresponding to each priority order in at least one decision scheme;

[0009] If it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

[0010] In a second aspect, an embodiment of the present application provides a decision-making device based on multi-objective optimization, including:

[0011] An acquisition module is used to acquire decision parameters of the multi-objective optimization problem to be solved, wherein the decision parameters at least include: multiple objective functions affecting the decision, and the priority order of the multiple objective functions;

[0012] A random generation module, used for randomly generating an initial decision population including multiple decision schemes based on multiple objective functions and priority orders of the multiple objective functions;

[0013] A construction module is used to construct an iterative decision population according to a priority characteristic value corresponding to at least one decision scheme in the initial decision population, wherein the priority characteristic value is a characteristic value obtained by performing non-dominated sorting on the objective function corresponding to each priority order in at least one decision scheme;

[0014] The determination module is used to determine the iterative decision population as the target decision population if it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations. The target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

[0015] In a third aspect, an embodiment of the present application provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any method of the first aspect when executing the computer program.

[0016] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method of any one of the first aspects is implemented.

[0017] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when executed on a terminal device, enables the terminal device to execute any one of the methods in the first aspect.

[0018] The embodiment of the present application provides a decision method, device and terminal device based on multi-objective optimization, the method includes: obtaining decision parameters of a multi-objective optimization problem to be solved, the decision parameters at least include: multiple objective functions affecting the decision, and the priority order of multiple objective functions; based on multiple objective functions and the priority order of multiple objective functions, randomly generating an initial decision population containing multiple decision schemes; according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, constructing an iterative decision population, wherein the priority characteristic value is the characteristic value obtained by non-dominated sorting of the objective functions corresponding to each type of priority order in at least one decision scheme; if it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved of the decision parameters. Using the above technical solution, by randomly generating an initial decision population according to the decision parameters of the multi-objective optimization problem to be solved, it is possible to perform iterative calculation according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, and obtain the target decision population used to solve the multi-objective optimization problem to be solved, so as to provide users with the best decision scheme that meets the priority requirements and improve the performance of the decision method. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 It is a flowchart of a decision-making method based on multi-objective optimization provided by an embodiment of the present application;

[0021] Figure 2 is a flowchart of a decision-making method based on multi-objective optimization provided by another embodiment of the present application;

[0022] Figure 3 It is a structural block diagram of a decision-making device based on multi-objective optimization provided by an embodiment of the present application;

[0023] Figure 4 It is a structural diagram of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0024] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0025] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0026] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0027] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0028] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0029] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0030] The decision-making method based on multi-objective optimization provided in the embodiments of the present application can be applied to terminal devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPC), netbooks, personal digital assistants (PDA), etc. The embodiments of the present application do not impose any restrictions on the specific types of terminal devices.

[0031] It can be considered that a multi-objective optimization problem refers to an optimization problem in which there are multiple conflicting or influencing objective functions. For example, the design of automobile body parts needs to consider both the stiffness and mass of the parts. These two objectives usually conflict with each other. For example, increasing stiffness will increase mass, while reducing mass will reduce stiffness. Therefore, the multi-objective evolutionary algorithm needs to make a trade-off between multiple conflicting objective functions and find an optimal solution that comprehensively considers multiple objective functions.

[0032] However, traditional multi-objective optimization problems do not take into account a widespread phenomenon in reality, that is, multiple objective functions participate in decision-making at different priority levels, and different objective functions have different importance to decision-making. Such multi-objective optimization problems are also called mixed Pareto-lexicographic multi-objective optimization problems with priority layers, and are widely present in many real-world problems such as vehicle design, post-disaster management, and industrial scheduling. For example, post-disaster management usually needs to consider multiple objectives, among which rapid rescue of disaster victims and prevention of secondary disasters during rescue are usually considered to be the two most important objectives, while restoring critical infrastructure and optimizing resource allocation are also important, but are usually given secondary priority.

[0033] When solving such multi-objective optimization problems with priorities, traditional multi-objective evolutionary algorithms usually assume that all objective functions have the same importance to the decision, which makes it impossible to efficiently provide the best solution that meets the user's priority requirements.

[0034] Based on this, an embodiment of the present application provides a decision-making method based on multi-objective optimization. By incorporating priority information into the mating mechanism of the multi-objective evolutionary algorithm, a new multi-priority multi-objective evolutionary algorithm is obtained, which can efficiently solve the mixed Pareto-lexicographic order multi-objective optimization problem and improve the user experience.

[0035] Figure 1 This is a flowchart of a decision-making method based on multi-objective optimization provided in an embodiment of the present application. As an example but not a limitation, the method can be applied to a terminal device.

[0036] S101. Obtain decision parameters of a multi-objective optimization problem to be solved, where the decision parameters at least include: multiple objective functions that affect the decision, and a priority order of the multiple objective functions.

[0037] The multi-objective optimization problem to be solved may refer to a multi-objective optimization problem with priorities that needs to be solved, and the decision parameters are parameters related to the multi-objective optimization problem to be solved, such as decision parameters that may include multiple objective functions that affect the decision, and the priority order of multiple objective functions, and may also include hyperparameters of the optimization problem, such as decision variables. Furthermore, in addition to obtaining the above-mentioned related content, algorithm parameters may also be input at the same time, such as the target number of decisions and / or the number of algorithm iterations, etc. The target number of decisions may refer to the number of decision solutions that need to be returned in the end, and the number of algorithm iterations is the number of rounds of iterative optimization calculations performed by the algorithm.

[0038] S102: Based on multiple objective functions and the priority order of the multiple objective functions, randomly generate an initial decision population including multiple decision solutions.

[0039] The initial decision population can be considered as a preliminarily generated decision population. The initial decision population contains multiple decision schemes, and subsequent iterative calculations are required to obtain the optimal decision scheme for users to select and use.

[0040] In this embodiment, the specific process of generating the initial decision population is not limited and can be determined according to the specific actual situation. For example, first, based on multiple objective functions and the priority order of the multiple objective functions, decision schemes for the target decision number can be randomly generated, and then the initial decision population can be constructed through the decision schemes for the target decision number. Exemplarily, the generated decision schemes for the target decision number can be directly used as the initial decision population, or the initial decision population can be obtained by further processing, such as by sorting the decision schemes for the target decision number to obtain the initial decision population.

[0041] S103. Construct an iterative decision population according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, wherein the priority characteristic value is the characteristic value obtained by non-dominated sorting of the objective function corresponding to each type of priority order in at least one decision scheme.

[0042] The iterative decision population can be understood as a decision population obtained by iteratively calculating the initial decision population, and the number of decision solutions in the iterative decision population is the same as the number of decision solutions in the initial decision population.

[0043] The priority eigenvalue can be understood as the eigenvalue obtained by non-dominated sorting of the objective function corresponding to each type of priority order in the decision scheme. The number of priority eigenvalues ​​can correspond to the number of priority levels. The multiple priority eigenvalues ​​corresponding to each decision scheme can constitute the arrangement number of each decision scheme in the initial decision population. The arrangement number can refer to the number of each decision scheme obtained after sorting all the decision schemes in the initial decision population using priority fast non-dominated sorting.

[0044] For example, the ranking order of decision plan x in the initial decision population can be expressed using the Grossone framework. r1, r2, …, r L It can be the priority characteristic value corresponding to the decision plan x. For example, r1 can represent the rank obtained by non-dominated sorting of multiple objective functions of the first priority, and r2 represents the rank obtained by further non-dominated sorting of multiple objectives of the second priority based on the previous sorting. Rank is a positive integer. ① It can be the benchmark of the framework, which is fixed; ① 0 ① -1 ,…,① -(L-1) Characterizes each priority, for example, ① 0 Indicates the first layer priority, ① -1 In the field of hyperreal numbers (non-standard analysis), "Grossone" can be used to represent a new infinite unit of quantity, which is introduced as a practical and operational infinite mathematical entity for more detailed treatment of mathematical problems involving infinity and infinite processes.

[0045] In this step, the iterative decision population can be constructed according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population to obtain a decision population for solving the multi-objective optimization problem to be solved for the decision parameters. The construction process of the specific iterative decision population is not limited here. For example, the iterative decision population can be initially constructed according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, and then the iterative decision population can be updated according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population or the iterative decision population, so as to obtain the constructed iterative decision population through multiple iterative calculations.

[0046] S104. If it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

[0047] In the process of obtaining the iterative decision population through the above iterative calculation, it can be detected whether the number of times the iterative decision population is constructed reaches the number of algorithm iterations to determine the subsequent decision steps.

[0048] In some embodiments, if it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, it means that the iteration end condition has been met. Then the final iterative decision population can be determined as the target decision population, so as to use the target decision population to solve the multi-objective optimization problem of decision parameters to be solved.

[0049] In some embodiments, if it is detected that the number of times the iterative decision population is constructed has not reached the number of algorithm iterations, it means that the iteration end condition has not been met and the iterative calculation of the iterative decision population needs to continue. In this case, the iterative decision population can be used as the initial decision population, and the step of constructing the iterative decision population according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population is returned until it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations.

[0050] The present embodiment provides a decision method based on multi-objective optimization, which obtains decision parameters of a multi-objective optimization problem to be solved, wherein the decision parameters at least include: multiple objective functions that affect the decision, and the priority order of the multiple objective functions; based on the multiple objective functions and the priority order of the multiple objective functions, randomly generates an initial decision population containing multiple decision schemes; according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, constructs an iterative decision population, wherein the priority characteristic value is the characteristic value obtained by non-dominated sorting of the objective functions corresponding to each type of priority order in at least one decision scheme; if it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved of the decision parameters. By using this method, by randomly generating an initial decision population according to the decision parameters of the multi-objective optimization problem to be solved, it is possible to perform iterative calculation according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, and obtain the target decision population used to solve the multi-objective optimization problem to be solved, thereby providing users with the best decision scheme that meets the priority requirements and improving the performance of the decision method.

[0051] Figure 2It is a flowchart of a decision-making method based on multi-objective optimization provided by another embodiment of the present application. This embodiment will further optimize the construction of an iterative decision population based on the priority characteristic value corresponding to at least one decision plan in the initial decision population as follows: randomly select a decision plan from the initial decision population as the first decision plan; determine the second decision plan based on the priority characteristic value corresponding to the first decision plan, and construct a descendant decision population based on the first decision plan and the second decision plan; return to execute the step of randomly selecting a decision plan from the initial decision population as the first decision plan, until the number of decision plans in the descendant decision population reaches a preset number of decisions, and obtains the constructed descendant decision population; based on the initial decision population and the constructed descendant decision population, generate an iterative decision population. Figure 2 As shown, the method includes:

[0052] S201. Obtain decision parameters of a multi-objective optimization problem to be solved, where the decision parameters at least include: multiple objective functions that affect the decision, and a priority order of the multiple objective functions.

[0053] S202: Based on multiple objective functions and the priority order of the multiple objective functions, randomly generate an initial decision population including multiple decision solutions.

[0054] S203. Randomly select a decision plan from the initial decision population as the first decision plan.

[0055] The first decision plan may refer to a decision plan randomly selected from the initial decision population, and the random selection method is not limited as long as the selection of the first decision plan can be achieved.

[0056] S204: Determine a second decision plan based on the priority characteristic value corresponding to the first decision plan, and construct a descendant decision population according to the first decision plan and the second decision plan.

[0057] The second decision plan can be considered as a decision plan in the initial decision population. The second decision plan can be the same as the first decision plan, or it can be a decision plan different from the first decision plan. It can be specifically determined by the strategy preset by relevant personnel. Different preset strategies can determine different second decision plans. For example, the corresponding second decision plan can be directly selected from the initial decision population according to the size of the priority characteristic value corresponding to the first decision plan, or the second decision plan can be determined through a series of calculations. This embodiment does not limit this.

[0058] In some embodiments, determining the second decision solution based on the priority feature value corresponding to the first decision solution includes:

[0059] According to the priority characteristic value and the number of priority layers corresponding to the first decision scheme, the priority position parameter value of the first decision scheme is calculated, where the priority position parameter value is the priority position corresponding to the first priority characteristic value greater than the set value among the priority characteristic values ​​corresponding to the first decision scheme, or the sum of the number of priority layers and the set value;

[0060] If it is detected that the priority position parameter value is less than or equal to the number of priority layers, the second decision plan is determined according to the priority characteristic value and the set function corresponding to each decision plan in the initial decision population, and the set function is a function constructed according to the priority characteristic value and the priority position parameter value corresponding to the first decision plan;

[0061] If it is detected that the priority position parameter value is greater than the priority layer number, a decision plan is randomly selected from the initial decision population as the second decision plan.

[0062] Among them, the number of priority levels can be determined according to the priority order of the multiple objective functions obtained. If there are several priority orders, the number of priority levels can be obtained accordingly. The priority position parameter value can be understood as a parameter value related to the priority position. For example, the priority position can be the priority position corresponding to the first priority feature value in the priority feature value that is greater than the set value, or the priority position can be the sum of the priority level and the set value. It can be specifically determined according to the priority feature value and the priority level corresponding to the first decision scheme.

[0063] For example, the priority position e s It can be expressed by the formula To determine specifically, L is the priority layer number, the set value is 1, when the cumulative value of the priority characteristic value corresponding to the first decision scheme is greater than the priority layer number L, the priority position can be the priority position corresponding to the first priority characteristic value in the priority characteristic value that is greater than the set value, otherwise, the priority position can be the sum of the priority layer number and the set value.

[0064] Furthermore, assuming that the first decision solution y is ranked in the initial decision population r1, r2, and r3 are 1, 2, and 1 respectively. And the priority position corresponding to the first priority eigenvalue greater than 1 in the priority eigenvalue is 2, that is, the priority position parameter value e s Equals 2.

[0065] The aggregate function may be a function constructed according to the priority characteristic value and the priority position parameter value corresponding to the first decision solution. The specific construction method may be determined according to the actual situation. For example, when the priority characteristic values ​​corresponding to the first decision solution y are r1, r2, ..., r L , the priority position parameter is es When

[0066] in, represents any positive integer greater than or equal to 1, and L is the number of priority levels.

[0067] In a specific implementation, the calculated priority position parameter value and the priority layer number can be detected to obtain a detection result, so as to determine a specific second decision plan based on the detection result. By way of example, if it is detected that the priority position parameter value is less than or equal to the priority layer number, the specific second decision plan can be further determined based on the priority characteristic value and set function corresponding to each decision plan in the initial decision population. This embodiment does not further elaborate on the means for specifically determining the second decision plan, as long as the second decision plan can be determined.

[0068] Correspondingly, if it is detected that the priority position parameter value is greater than the priority layer number, then a decision plan can be randomly selected from the initial decision population as the second decision plan.

[0069] Furthermore, the offspring decision population in this embodiment can be another population generated by the algorithm, which is used to generate a corresponding iterative decision population in combination with the initial decision population. The specific method of generating the offspring decision population in this embodiment can be iterative generation. Assuming that the offspring decision population is an empty set, the offspring decision population can be constructed step by step by repeatedly determining the first decision scheme and the second decision scheme.

[0070] In some embodiments, constructing a descendant decision population according to the first decision scheme and the second decision scheme includes:

[0071] Performing cross-mutation on the first decision plan and the second decision plan to obtain a new first decision plan and a new second decision plan;

[0072] Add the new first decision plan and the new second decision plan to the offspring decision population.

[0073] In a specific implementation, after the first decision scheme and the second decision scheme are determined by the above steps, the first decision scheme and the second decision scheme can be cross-mutated to obtain a new first decision scheme and a new second decision scheme, and then the new first decision scheme and the new second decision scheme obtained in this round can be added to the offspring decision population, and the above operation is repeated until the number of decision schemes in the offspring decision population reaches the preset number of decisions, that is, the step of randomly selecting a decision scheme from the initial decision population as the first decision scheme can be returned to perform, and the second decision scheme of the next construction cycle is determined according to the priority characteristic value corresponding to the first decision scheme selected in the next construction cycle and cross-mutated again, so as to add the new first decision scheme and the new second decision scheme obtained by mutation to the offspring decision population. The preset number of decisions can be configured in advance by relevant personnel or users, such as obtaining the corresponding preset number of decisions while obtaining the decision parameters.

[0074] S205, return to execute the step of randomly selecting a decision plan from the initial decision population as the first decision plan, until the number of decision plans in the offspring decision population reaches a preset number of decisions, and the constructed offspring decision population is obtained.

[0075] S206: Generate an iterative decision population based on the initial decision population and the constructed offspring decision population.

[0076] This step can generate an iterative decision population based on the initial decision population and the constructed offspring decision population. For example, the initial decision population and the constructed offspring decision population can be directly merged to obtain the iterative decision population. The iterative decision population can also be constructed by selecting a decision scheme with a preset number of decisions. The selection rules are not limited and can be determined according to actual needs.

[0077] Exemplarily, the process of generating an iterative decision population is, for example, to merge the initial decision population and the constructed offspring decision population to obtain a merged decision population, and to sort the decision solutions in the merged decision population using a non-dominated sorting method with priority and a crowding distance mechanism with priority, and finally to select a preset number of decision solutions from the processed decision population as the iterative decision population, such as selecting the top N decision solutions as the iterative decision population. Among them, the number of decision solutions in the initial decision population, the constructed offspring decision population, and the iterative decision population are all preset decision numbers.

[0078] S207: If it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

[0079] The present embodiment provides a decision-making method based on multi-objective optimization. By determining the second decision plan based on the priority characteristic value corresponding to the selected first decision plan, and constructing a child decision population by iterative calculation, it can provide a feasible way for generating an iterative decision population, and further better generate a target decision population for solving the multi-objective optimization problem to be solved, thereby improving the performance of the decision-making method.

[0080] In some embodiments, determining the second decision solution according to the priority characteristic value and the set function corresponding to each decision solution in the initial decision population includes:

[0081] For each decision plan in the initial decision population, determine whether the priority eigenvalue corresponding to each decision plan is a solution of the aggregate function;

[0082] Constructing at least one decision solution corresponding to a solution of the set function whose priority characteristic value is a decision solution set;

[0083] A decision plan is randomly selected from the decision plan set as the second decision plan.

[0084] The decision solution set may refer to a set of several decision solutions, and the specific content may be determined according to whether the priority characteristic value corresponding to the decision solution is a solution of the set function.

[0085] In this embodiment, in the process of determining the second decision plan according to the priority characteristic value and the set function corresponding to each decision plan in the initial decision population, it can be judged whether the arrangement sequence number of each decision plan in the initial decision population satisfies the set function, that is, for each decision plan in the initial decision population, it can be judged whether the priority characteristic value corresponding to each decision plan is a solution of the set function; if the priority characteristic value corresponding to a decision plan is a solution of the set function, this decision plan is added to the decision plan set, so that after the decision plan set is constructed, a decision plan can be randomly selected from the decision plan set as the second decision plan.

[0086] The following is an exemplary description of a decision-making method based on multi-objective optimization provided in this embodiment. First, a hybrid Pareto-lexicographic order multi-objective optimization problem with a priority layer can be modeled, and the multi-priority multi-objective evolutionary algorithm provided in this embodiment can be used to solve it, and finally a set of different optimal decision solutions can be obtained.

[0087] For example, a hybrid Pareto-lexicographic multi-objective optimization problem with L-level priorities can be described as: Among them, the i-th layer priority contains m i objective functions, a total of One objective function. The importance of the objective functions in the first-level priority is much higher than that of the objective functions in the second-level priority, and so on; multiple objective functions within the same level have the same importance.

[0088] Next, parameters such as the problem description, population size N (i.e., the preset number of decisions), and the maximum number of fitness evaluations FEmax (i.e., the number of algorithm iterations) can be input into the multi-priority multi-objective evolutionary algorithm. The algorithm can first randomly generate N decision-making schemes as the initial population, denoted as P; then start the iterative process: set e = N, and when e < FEmax, perform the following operations:

[0089] Call the new parent selection operator to generate N new decision-making schemes (i.e., the offspring population) based on the population P, and denote it as Q. For example, 2 decision-making schemes can be taken out from the current population each time for crossover and mutation to generate 2 offspring decision-making schemes, and finally form the offspring population (the specific process can be seen in the subsequent description); merge the current population P and the offspring population Q, and use the non-dominated sorting method with priority and the crowding distance mechanism with priority to sort each decision-making scheme in the merged population to obtain the population R; set P as the top N best decision-making schemes in R; set e = e + N, and return to the step of calling the new parent selection operator to generate N new decision-making schemes based on the population P until e ≥ FEmax, and finally return the population P.

[0090] Next, the new parent selection operator is introduced. Specifically, first, the offspring population Q can be initialized as an empty set, and then enter the iterative process: let i = 1, and when i ≤ N / 2, perform the following operations:

[0091] Randomly select a decision-making scheme from the current population P, denoted as x, and calculate the priority position parameter value e of the decision-making scheme x by reading the arrangement serial number of the decision-making scheme x s ; construct a set function based on the priority characteristic value and the priority position parameter value of the decision-making scheme x, and select the solutions of the decision-making schemes in the current population P that satisfy the set function to form a decision-making scheme set denoted as MP(x).

[0092] Next, if e s ≤ L, randomly select a decision-making scheme x' from the set MP(x); otherwise, randomly select a decision-making scheme x' from the current population P; cross and mutate x and x' to generate two new decision-making schemes, and add the two new decision-making schemes to the offspring population Q, set i = i + 1, and return to the step of randomly selecting the decision-making scheme x from the current population P until i > N / 2, and finally return the offspring population Q.

[0093] Taking post-disaster management as an example, the decision-making method based on multi-objective optimization provided by this embodiment is further explained. For example, the problem model may include: decision variables, multiple objective functions that affect the decision, and the priority order of multiple objective functions. Assuming there are 4 disaster-stricken points, how many rescue personnel should be deployed at each disaster-stricken point (can be set to x1, x2, x3, x4); how much rescue supplies should be deployed at each disaster-stricken point (can be set to x5, x6, x7, x8), there will be 4*2=8 decision variables.

[0094] The objective function and priority order may include: minimizing the final medical losses of all disaster-affected sites (priority order is 1); minimizing the time required to complete the rescue of all disaster-affected sites (priority order is 1); minimizing the final economic losses of all disaster-affected sites (priority order is 2); minimizing other expenses required to complete the rescue of all disaster-affected sites (priority order is 2).

[0095] The problem model and algorithm parameters (such as the number of rescue solutions N returned in the end and the number of optimization iterations) can be input into the improved algorithm. The final algorithm can output N different (x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14, x15, x16, x17, x18, x19, x20, x21, x30, x40, x d , x7, x8), where each (x1, x2, x3, x4, x5, x6, x7, x8) represents a rescue plan. For example, (400, 200, 300, 100, 20, 10, 50, 80) can mean that the first disaster site corresponds to 400 rescuers and 200,000 rescue supplies; the second disaster site corresponds to 200 rescuers and 100,000 rescue supplies; the third disaster site corresponds to 300 rescuers and 500,000 rescue supplies; the fourth disaster site corresponds to 100 rescuers and 800,000 rescue supplies.

[0096] From the above description, it can be found that the decision-making method based on multi-objective optimization provided in this embodiment obtains an efficient multi-priority multi-objective evolutionary algorithm by incorporating the user's priority requirements for different objective functions into the mating mechanism of the multi-objective evolutionary algorithm. It can efficiently and effectively solve the mixed Pareto-lexicographic order multi-objective optimization problem with priority layers, and provide a set of diversified decision-making solutions that meet the user's priority requirements.

[0097] Specifically, the algorithm includes a hybrid Pareto-Lexicographic mating selection strategy based on niche technology to enhance the search within each priority level, and then integrates new components into the conventional multi-priority multi-objective evolutionary algorithm to form a new multi-priority multi-objective evolutionary algorithm, namely, the mixed Pareto-Lexicographic Evolutionary Algorithm with HierArchical Matingselection (PL-HAM-EA).

[0098] Corresponding to the decision-making method based on multi-objective optimization in the above embodiment, Figure 3 This is a structural block diagram of a decision-making device based on multi-objective optimization provided in one embodiment of the present application. For the sake of ease of explanation, only the parts related to the embodiment of the present application are shown.

[0099] Reference Figure 3 , the device comprises:

[0100] An acquisition module 301 is used to acquire decision parameters of a multi-objective optimization problem to be solved, wherein the decision parameters at least include: multiple objective functions that affect the decision, and the priority order of the multiple objective functions;

[0101] A random generation module 302, used to randomly generate an initial decision population including multiple decision solutions based on multiple objective functions and the priority order of the multiple objective functions;

[0102] A construction module 303 is used to construct an iterative decision population according to a priority characteristic value corresponding to at least one decision scheme in the initial decision population, wherein the priority characteristic value is a characteristic value obtained by performing non-dominated sorting on the objective function corresponding to each priority order in at least one decision scheme;

[0103] The determination module 304 is used to determine the iterative decision population as the target decision population if it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations. The target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

[0104] The present embodiment provides a decision-making device based on multi-objective optimization, which obtains decision parameters of a multi-objective optimization problem to be solved through an acquisition module, and the decision parameters at least include: multiple objective functions affecting the decision, and the priority order of multiple objective functions; the random generation module randomly generates an initial decision population containing multiple decision schemes based on multiple objective functions and the priority order of multiple objective functions; the construction module constructs an iterative decision population according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, wherein the priority characteristic value is the characteristic value obtained by non-dominated sorting of the objective functions corresponding to each type of priority order in at least one decision scheme; if the determination module detects that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved of the decision parameters. By using the device, by randomly generating an initial decision population according to the decision parameters of the multi-objective optimization problem to be solved, it is possible to perform iterative calculation according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, and obtain the target decision population used to solve the multi-objective optimization problem to be solved, thereby providing users with the best decision scheme that meets the priority requirements and improving the performance of the decision method.

[0105] Optionally, the decision-making device based on multi-objective optimization provided in this embodiment further includes:

[0106] The execution module is returned, which is used to use the iterative decision population as the initial decision population if it is detected that the number of times the iterative decision population is constructed does not reach the number of algorithm iterations, and return to execute the step of constructing the iterative decision population according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, until it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations.

[0107] Optionally, building blocks include:

[0108] A random selection unit, used for randomly selecting a decision plan from the initial decision population as a first decision plan;

[0109] A determination unit, configured to determine a second decision scheme based on a priority characteristic value corresponding to the first decision scheme, and to construct a descendant decision population according to the first decision scheme and the second decision scheme;

[0110] A return execution unit is used to return to the step of randomly selecting a decision scheme from the initial decision population as the first decision scheme, until the number of decision schemes in the descendant decision population reaches a preset number of decisions, thereby obtaining a constructed descendant decision population;

[0111] The generation unit is used to generate an iterative decision population based on the initial decision population and the constructed offspring decision population.

[0112] Optionally, the determining unit includes:

[0113] A calculation subunit, used to calculate a priority position parameter value of the first decision scheme according to the priority characteristic value and the number of priority layers corresponding to the first decision scheme, wherein the priority position parameter value is the priority position corresponding to the first priority characteristic value greater than a set value among the priority characteristic values ​​corresponding to the first decision scheme, or the sum of the number of priority layers and the set value;

[0114] A determination subunit is used to determine a second decision scheme according to the priority characteristic value and the set function corresponding to each decision scheme in the initial decision population if it is detected that the priority position parameter value is less than or equal to the priority layer number, where the set function is a function constructed according to the priority characteristic value and the priority position parameter value corresponding to the first decision scheme;

[0115] The randomly selected subunit is used to randomly select a decision plan from the initial decision population as the second decision plan if it is detected that the priority position parameter value is greater than the priority layer number.

[0116] Optionally, determine that the subunit is specifically used for:

[0117] For each decision plan in the initial decision population, determine whether the priority eigenvalue corresponding to each decision plan is a solution of the aggregate function;

[0118] Constructing at least one decision solution corresponding to a solution of the set function whose priority characteristic value is a decision solution set;

[0119] A decision plan is randomly selected from the decision plan set as the second decision plan.

[0120] Optionally, the determination unit is specifically used for:

[0121] Performing cross-mutation on the first decision plan and the second decision plan to obtain a new first decision plan and a new second decision plan;

[0122] Add the new first decision plan and the new second decision plan to the offspring decision population.

[0123] Optionally, the generation unit is specifically used for:

[0124] Merge the initial decision population and the constructed offspring decision population to obtain a merged decision population;

[0125] Sorting the decision plans in the merged decision population;

[0126] A preset number of decision schemes are selected from the processed decision population as the iterative decision population, wherein the number of decision schemes in the initial decision population, the constructed offspring decision population and the iterative decision population are all the preset number of decisions.

[0127] Optionally, the random generation module is specifically used to:

[0128] Based on multiple objective functions and the priority order of multiple objective functions, randomly generate decision plans with a target number of decisions;

[0129] The decision plans for the target number of decisions are sorted to obtain the initial decision population.

[0130] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0131] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0132] The present application also provides a terminal device, Figure 4 is a schematic diagram of the structure of a terminal device provided by an embodiment of the present application, such as Figure 4 As shown, the terminal device includes: at least one processor 401, a memory 402, an input device 403, an output device 404, and a computer program stored in the memory 402 and executable on at least one processor 401. When the processor 401 executes the computer program, the steps in any of the above-mentioned method embodiments are implemented.

[0133] The input device 403 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the terminal device. The output device 404 may include a display device such as a display screen.

[0134] The embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by the processor 401, the steps in the above-mentioned method embodiments can be implemented.

[0135] An embodiment of the present application provides a computer program product. When the computer program product runs on a mobile terminal, the mobile terminal can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0136] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can be executed by the processor 401. The steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include at least: any entity or device that can carry the computer program code to the device / terminal device, a recording medium, a computer memory, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), an electric carrier signal, a telecommunication signal, and a software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0137] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0138] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0139] In the embodiments provided in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0140] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0141] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A decision-making method based on multi-objective optimization, characterized in that: include: Obtaining decision parameters of a multi-objective optimization problem to be solved, wherein the decision parameters include at least: a plurality of objective functions affecting the decision, and a priority order of the plurality of objective functions; Based on the multiple objective functions and the priority order of the multiple objective functions, randomly generate an initial decision population including multiple decision solutions; According to the priority characteristic value corresponding to at least one decision scheme in the initial decision population, an iterative decision population is constructed, wherein the priority characteristic value is a characteristic value obtained by performing non-dominated sorting on the objective function corresponding to each priority order in the at least one decision scheme; If it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations, the iterative decision population is determined as the target decision population, and the target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

2. The decision-making method based on multi-objective optimization according to claim 1, characterized in that: The method further comprises: If it is detected that the number of times the iterative decision population is constructed does not reach the number of algorithm iterations, the iterative decision population is used as the initial decision population, and the step of constructing an iterative decision population according to the priority characteristic value corresponding to at least one decision scheme in the initial decision population is returned to execute until it is detected that the number of times the iterative decision population is constructed reaches the number of algorithm iterations.

3. The decision-making method based on multi-objective optimization according to claim 1, characterized in that: The step of constructing an iterative decision population according to a priority characteristic value corresponding to at least one decision scheme in the initial decision population includes: Randomly select a decision plan from the initial decision population as the first decision plan; Determine a second decision scheme based on the priority characteristic value corresponding to the first decision scheme, and construct a descendant decision population according to the first decision scheme and the second decision scheme; Return to the step of randomly selecting a decision solution from the initial decision population as the first decision solution, until the number of decision solutions in the offspring decision population reaches a preset number of decisions, thereby obtaining a constructed offspring decision population; An iterative decision population is generated based on the initial decision population and the constructed offspring decision population.

4. The decision-making method based on multi-objective optimization according to claim 3, characterized in that: The determining of the second decision solution based on the priority characteristic value corresponding to the first decision solution includes: Calculate the priority position parameter value of the first decision scheme according to the priority characteristic value and the priority level number corresponding to the first decision scheme, wherein the priority position parameter value is the priority position corresponding to the first priority characteristic value greater than a set value among the priority characteristic values ​​corresponding to the first decision scheme, or the sum of the priority level number and the set value; If it is detected that the priority position parameter value is less than or equal to the priority layer number, a second decision scheme is determined according to the priority characteristic value and the set function corresponding to each decision scheme in the initial decision population, where the set function is a function constructed according to the priority characteristic value corresponding to the first decision scheme and the priority position parameter value; If it is detected that the priority position parameter value is greater than the priority level number, a decision plan is randomly selected from the initial decision population as the second decision plan.

5. The decision-making method based on multi-objective optimization according to claim 4, characterized in that: The step of determining the second decision solution according to the priority characteristic value and the set function corresponding to each decision solution in the initial decision population includes: For each decision scheme in the initial decision population, determining whether the priority characteristic value corresponding to each decision scheme is a solution of the set function; Constructing at least one decision solution corresponding to a solution of the set function whose priority characteristic value is a decision solution set; A decision plan is randomly selected from the decision plan set as the second decision plan.

6. The decision-making method based on multi-objective optimization according to claim 3, characterized in that: The step of constructing a descendant decision population according to the first decision scheme and the second decision scheme includes: Performing crossover mutation on the first decision plan and the second decision plan to obtain a new first decision plan and a new second decision plan; The new first decision plan and the new second decision plan are added to the offspring decision population.

7. The decision-making method based on multi-objective optimization according to claim 3, characterized in that: The step of generating an iterative decision population based on the initial decision population and the constructed offspring decision population includes: Merging the initial decision population and the constructed offspring decision population to obtain a merged decision population; Sorting the decision solutions in the merged decision population; A preset number of decision schemes are selected from the processed decision population as an iterative decision population, wherein the number of decision schemes in the initial decision population, the constructed offspring decision population and the iterative decision population are all the preset number of decisions.

8. The decision-making method based on multi-objective optimization according to claim 7, characterized in that: The step of randomly generating an initial decision population including a plurality of decision solutions based on the plurality of objective functions and the priority order of the plurality of objective functions comprises: Based on the multiple objective functions and the priority order of the multiple objective functions, randomly generating decision solutions of a target number of decisions; The decision schemes of the target decision number are sorted to obtain the initial decision population.

9. A decision-making device based on multi-objective optimization, characterized in that: include: An acquisition module, used to acquire decision parameters of a multi-objective optimization problem to be solved, wherein the decision parameters at least include: a plurality of objective functions affecting the decision, and a priority order of the plurality of objective functions; A random generation module, used for randomly generating an initial decision population including a plurality of decision solutions based on the plurality of objective functions and the priority order of the plurality of objective functions; A construction module, used to construct an iterative decision population according to a priority characteristic value corresponding to at least one decision scheme in the initial decision population, wherein the priority characteristic value is a characteristic value obtained by performing non-dominated sorting on the objective function corresponding to each priority order in the at least one decision scheme; A determination module is used to determine the iterative decision population as a target decision population if it is detected that the number of constructions of the iterative decision population reaches the number of algorithm iterations. The target decision population is used to solve the multi-objective optimization problem to be solved for the decision parameters.

10. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

Images