Improved Chicken Swarm Algorithm Based On Exploration And Exploitation Balance And Its Application

Chicken Swarm Optimization(CSO)is a novel swarm intelligence algorithm proposed in recent years,which simulates the foraging behavior of chickens in a hierarchical system in nature.The CSO algorithm divides the whole swarm into roosters,hens and chicks,in which the rooster leads the hen and the chicks follow the hen.Like other swarm intelligence algorithms,CSO algorithm also has some shortcomings,such as slow convergence speed and easy to fall into local optimal.For swarm intelligence algorithm,balancing exploration and exploitation in search process is an effective measure to improve algorithm performance.Therefore,this dissertation proposes a series of improved algorithms based on the search characteristics of CSO algorithm from the perspective of exploration-exploitation balance.The specific research work is as follows:Firstly,an improved CSO algorithm based on stimulus-response mechanism is proposed by using the balance exploration and exploitation of leading role.Specifically,considering the leading role of rooster,two search equations are designed to explore and exploit,and the stimulus and threshold values of executing the search equation were defined according to the population aggregation degree and the average improvement degree.Under the stimulusresponse mechanism,roosters fully play a leading role and lead the chickens to complete exploration and exploitation.Simulation experiments on the benchmark function optimization and the survival risk prediction model optimization of esophageal cancer show that the proposed algorithm is significantly better than other comparison algorithms.Secondly,an improved CSO algorithm based on enhanced dominance relationship is proposed by using hierarchical balance exploration and exploitation.Specifically,the roosterhen dominance relationship was established according to fitness,and the hens improved their exploitation capacity by preferentially following the rooster with high fitness.The hen-chick dominance relationship was established based on diversity,and the chicks improved their ability to explore by prioritising the diversity of the hens.In a rooster-hen-chick hierarchy,the swarm combine exploration and exploitation.Compared with other algorithms,benchmark function optimization and survival risk prediction model optimization of esophageal cancer are performed,and the results show the effectiveness of the proposed algorithm.Finally,an improved CSO algorithm based on diversity guidance is proposed by using population state balance exploration and exploitation.Specifically,the population diversity is calculated by the individual-center distance,and exploration and exploitation are divided according to the diversity.In the process of exploration,the promising solution is searched by using random solution information.In the exploitation process,the convergence rate is accelerated by using the information of optimal solution.Guided by diversity,the swarm alternates between exploration and exploitation.The experimental results show that the proposed algorithm is superior to or at least comparable to other algorithms in most cases.