Improvement And Application Of Chicken Swarm Optimization Algorithm

In recent years,many swarm intelligent algorithms have been invented and improved,and successfully applied to a variety of numerical and combinatorial optimization problems,such as differential evolution algorithm,particle swarm optimization algorithm,bat algorithm,and wolf pack algorithm.Since the swarm intelligence optimization algorithm can provide acceptable solutions for nonlinear,high-dimensional complex and NP hard problems within a reasonable time,its popularity will continue to increase.The chicken swarm optimization algorithm is a new swarm intelligence optimization algorithm proposed in 2014.The algorithm has the advantages of strong global search capability,self-adaptive capability,and multi-subgroup cooperative search capability,and can be widely applied to various optimizations problem.In this paper,the following researches are carried out for the improvement and application of the chicken optimization algorithm.(1)To overcome the drawbacks that the chicken swarm algorithm is liable to fall into the local optimum and is slow when solving the high-dimension problems,this paper proposes a hybrid chicken swarm algorithm with dissipative structure and differential mutation.According to the principle of the chicken swarm optimization algorithm,the cock leads the chickens in each subgroup to search for food,the cock affects the search direction and search speed of the entire chicken swarm.Therefore,the dissipative structure is adapted into the cock position to enlarge the search space of the whole flock and enhance the global searching ability.In addition,in the later stages of evolution,no effective information transfer between individuals can be achieved.Therefore,differential mutation is applied in some randomly selected individuals to suppress the rapid decline of population diversity and to enhance the convergence performance of the algorithm.By analyzing the improved algorithm and other algorithms from the aspects of convergence speed,convergence accuracy and stability,with 18 classical test functions.The simulation results show that the improved algorithm is effective and feasible.(2)To overcome the drawbacks that the chicken swarm algorithm is low convergence speed and low precision when solving the 0-1 knapsack problem,an improved binary chicken swarm optimization algorithm is proposed.In order to maintain the excellent evolutionary characteristics of the basic chicken swarm optimization algorithm,a real and binary mixed code is proposed,that is the algorithm work on a real vector,and the solution is represented by binary vector.In addition,a repair strategy is proposed to enhanced the quality of the solution.Finally,a mutation operation is designed to reduce the risk of the algorithm falling into a local optimum and to accelerate the algorithm's ability to search for space.The improved algorithm and other typical algorithms are used to simulate 10 knapsack problems.The convergence speed,convergence accuracy,stability and running time are used as the four criteria.The results show that the improved chicken swarm optimization algorithm is solving the 0-1 backpack problem is more effective.