| With the continuous development of market economy,investment decision-making behavior related to economic benefits is increasingly complex,investors need to face with a series of complicated decision-making and optimization problems during market investment.Evolutionary computation is a kind of computational intelligence method that simulates the evolutionary mechanism of natural organisms,which is an important branch of artificial intelligence research.Evolutionary computation methods have highly self-organizing,selfadaptive,and global-searching capabilities,it is currently an important tool for solving complex optimization problems and one of the effective techniques to solve complex market economic benefit optimization problem.Especially with the development of multi-objective evolutionary computation methods,they can simultaneously consider multiple optimization objectives and provide a set of approximate Pareto optimal solutions that balance multiple optimization objectives for decision-makers.Therefore,evolutionary computation methods have broad application prospects in market economic benefit optimization problem.However,due to the complexity of market investment decision-making behavior,the multi-objective evolutionary computation methods still face the following challenges: Firstly,investment decision-making behavior often exists a variety of constraints,it often has both continuous equality or inequality constraints and discrete category or cardinality constraints,these constraints challenge the multi-constraint processing capability of the multi-objective evolutionary computation methods.Secondly,investment decision-making behavior often has multiple decision-makers,their decision-making goals are usually in conflict.When consultation and cooperation can be carried out between decision-makers,it is necessary to fully consider the respective optimization objectives and decision-making preferences of participants in negotiation.There may be multiple preferences that need to be comprehensively considered.The relevant mechanism to make the multi-objective evolutionary computation methods simultaneously consider the optimization objectives and multiple preferences of decision-makers still needs further study.Thirdly,when consultation and cooperation cannot be carried out between decision-makers,each decision-maker will preferentially optimize its own goal,they are gaming with each other rather than cooperating with each other.At this moment,how to guide the decision-maker’s behavior from the global perspective through government departments and let the decision-makers present a collaboration towards the overall optimal goal of the market,this is a new optimization problem which needs further study.To solve the above challenges,this thesis conducts research on multi-objective evolutionary computation methods involving investment decision-making behavior in complex market economic benefit optimization.Research is carried out from the perspectives of singleparticipants multi-constraint,dual-participants multi-preference,and multi-participants noncooperative:(1)To meet the multi-objective optimization requirements involving single-participants multi-constraint in market economic benefit optimization,a knowledge-based constructive estimation of distribution algorithm is designed for solving the portfolio optimization problem with cardinality constraints.Firstly,a hybrid design of ant colony optimization and estimation of distribution algorithm is used to solve this mixed-variable optimization problem based on knowledge information.Second,a knowledge accumulation mechanism is designed to discover the internal relationship among the assets.The mechanism can not only guide the selection of assets effectively but also enable the use of historical information during evolution to direct the allocation of investment proportion.Third,a constructive approach is applied to construct portfolios under discrete and continuous constraints.This method can solve the multi-constraint problem based on knowledge information and construct an efficient multi-objective optimal solution set under complex constraints.(2)To meet the multi-objective optimization requirements involving dual-participants multi-preference in market economic benefit optimization,a multi-level region interest preference-based multi-objective evolutionary algorithm is designed for solving the payment scheduling negotiation problem.Based on the multi-objective and multi-preference characteristic of this problem,the problem is reformulated as a bi-objective optimization problem with preferences at first.Second,to address the different preferences of the client and the contractor,a strategy of multi-level region interest is presented.Third,this strategy is integrated in the non-dominated sorting genetic algorithm II to solve the payment scheduling negotiation problem efficiently.This method can solve the multi-preference integration problem,the proposed method can focus on searching in the region of interest and provide more satisfactory solutions.(3)To meet the multi-objective optimization requirements involving multi-participants non-cooperative in market economic benefit optimization,a multi-participants adaptive guidance evolutionary computation method is designed for solving the vehicle energy station distribution problem.From the individual perspective,we use a network-based evolutionary game with a confidence mechanism to describe the behavior of investors.From the global perspective,we design global satisfaction and social benefit indicators to evaluate the vehicle energy station distribution program.From the government perspective,we design two individual guidance method for the government to motivate selfish investors to adopt strategies in accord with the overall interests of all customers by using subsidy and taxation.This method can solve the multi-participants guidance problem,the proposed method can effectively guide the individuals to make decisions that conducive to the global objective.In summary,the research carried out in this thesis on the multi-objective evolutionary computation method for complex market economic benefit optimization problem involving investment decision-making behavior,mainly aimed to improve the multi-constraint processing,multi-preference integration,and multi-participants guidance capabilities of multi-objective evolutionary methods.This research can also enhance the investment decision-making optimization performance of complex market economic benefit problems. |
