Research On Multi-Swarm Multi-Strategy Particle Swarm Algorithm And Portfolio Problem

Swarm intelligence is a general term for biological groups that can exhibit intelligent behavior.The imitation of the behavior of these biological groups inspired the birth of swarm intelligence algorithms.The swarm intelligence optimization algorithm can always play an unexpected and advantageous role in the face of large-scale optimization problems,so it is considered to be a very valuable research direction.The particle swarm optimization algorithm is an important branch of the swarm intelligence optimization algorithm.It is an analogous modeling for the foraging activities of birds.In the iterative process of the algorithm,birds are regarded as particles,and some evolution strategies are used to update the particles to simulate the flight search of birds.The idea of particle swarm optimization is simple and easy to implement,but prone to fall into the local optima and of low convergence accuracy.In order to improve the performance of the PSO,a lot of improvement work have been done,and found that the introduction of multiple swarm operations and different perturbation strategies can effectively improve the problem of low convergence accuracy of particle swarm optimization.This paper introduces the standard particle swarm optimization algorithm and the multi-swarm multi-learning strategy idea in detail,and according to this idea,proposes a multi-strategy comprehensive learning particle swarm optimization algorithm(MSPSO)based on population partitioning.The algorithm uses the competition mechanism to divide the population into two subgroups:potential subgroup and ordinary subgroup.Different evolution strategies are implemented for these two subgroups.The particles in the potential subgroup are mainly responsible for global exploration,while the particles in the ordinary subgroup are focus on local exploration.Compared with other excellent swarm intelligence algorithms on different types of benchmark functions,the proposed algorithm MSPSO shows the best effect,which proves that the proposed algorithm has better algorithm performance.Portfolio optimization is the process of selecting the best portfolio from all the portfolios under consideration based on a certain objective.The goal is usually to achieve higher returns or minimize risk.It is an extremely complex optimization problem,which is difficult for traditional algorithms to deal with.In this paper,the proposed MSPSO algorithm is used to solve the investment portfolio problem,and the effectiveness of the proposed algorithm is verified,which shows that the swarm intelligence algorithm can be very suitable for solving practical problems.