Research On Multi-Objective Two-Sided Matching Problem Based On Improved Particle Swarm Ant Colony Algorithm

In recent years, as the swift and violent development of electronic commerce in our country, electronic brokers in intermediaries will be as much as possible to meet the demand conditions of both buyers and sellers and the interests of all parties to facilitate the transactions in electronic commerce transaction process, which is called two-sided matching problem on electronic commerce. There are a lot of two-sided matching problems in real life. At present, most of the algorithms for solving multi-objective two-sided matching problem are to transform it into a single objective problem. However, different decision makers consider the same problem from different perspectives. Therefore, this paper establishes a multi-objective two-sided matching problem model, and applying an improved particle swarm ant colony algorithm to solve it.The main work is as follows:(1)A global optimal mutation strategy is proposed to improve the search efficiency of particle swarm algorithm in continuous optimization problem, which is latterly integrated into the adaptive particle swarm optimization algorithm with shrink and expansion operation to keep the whole particle group has high search efficient and the ability to escape from the local optimum in time. By testing several groups of classical test functions, the experiment shows that this method can enhance the search performance and improve the accuracy and speed of the search progress.(2)A fully adaptive pheromone adjustment on multi-objective ant colony algorithm is proposed direct against colony problem of low convergence speed and easily falling into local optimum. Through dynamically controlling pheromone information of ant colony system, the algorithm can be capable of faster convergence and getting rid of stagnation, and the multi-objective non-dominated solution has good convergence and uniform distribution. Then having several group tests with the internationally recognized TSP multi-objective test functions, analyze the results.(3)A method of calculating satisfaction degree of the two-sided matching problem is designed, which can reflect the degree of satisfaction more accurately. Then a kind of mathematical model of two-sided matching problem is designed based on it. For multi-objective two-sided matching problem characteristics, the particle swarm algorithm and the ant colony algorithm of (1)(2) were improved separately, then combined organically to be applied to the established mathematical model of two-sided matching problem. The relevant experimental results show that the improved multi-objective particle swarm ant colony algorithm can aptly solve this kind of two-sided matching problem.