CAJ | 학술논문

粒子群算法是一种新的群体智能算法,被广泛用于各种复杂优化问题的求解,但算法存在着过早收敛问题.为了克服算法早熟的缺点,将粒子群看作是一个复杂的免疫系统,借鉴生物学中免疫系统自我调节的机制,提出了一种新的基于免疫选择的粒子群优化算法(IS-PSO).免疫系统中的抗原、抗体和亲和度分别对应了待优化函数的最优解、候选解和适应度.IS-PSO通过免疫算法中免疫记忆、疫苗接种、免疫选择等操作有效地调节PSO算法中种群的多样性.给出了算法的详细步骤,并将本文提出的算法与基本的粒子群算法(bPSO)在几个典型Benchmark函数的优化问题应用中进行了比较,仿真结果表明:IS-PSO算法可以有效避免早熟问题,提高粒子群算法求解复杂函数的全局优化性能.
입자군산법시일충신적군체지능산법,피엄범용우각충복잡우화문제적구해,단산법존재착과조수렴문제.위료극복산법조숙적결점,장입자군간작시일개복잡적면역계통,차감생물학중면역계통자아조절적궤제,제출료일충신적기우면역선택적입자군우화산법(IS-PSO).면역계통중적항원、항체화친화도분별대응료대우화함수적최우해、후선해화괄응도.IS-PSO통과면역산법중면역기억、역묘접충、면역선택등조작유효지조절PSO산법중충군적다양성.급출료산법적상세보취,병장본문제출적산법여기본적입자군산법(bPSO)재궤개전형Benchmark함수적우화문제응용중진행료비교,방진결과표명:IS-PSO산법가이유효피면조숙문제,제고입자군산법구해복잡함수적전국우화성능.
Particle swarm optimization (PSO), a novel swarm intelligence algorithm, is proved to be a valid optimization technique and has been applied in many areas successfully. However, like other evolutionary algorithms, PSO also suffered from the premature convergence problem, especially for the large scale and complex problems. In order to overcome the shortcoming, this paper regards the swarm as a complex immune system, uses for reference from the self-adjustment mechanism of immune system, and proposes a novel PSO based on immune selection called IS-PSO(immune selection particle swarm optimization). Antigen, antibody, and affinity between antigen and antibody are corresponding to the best solution, the candidate solution, and the fitness values of the solution on the objective function, respectively. IS-PSO adjusts swarm diversity via immune memory, inoculate vaccine and immune selection and so on. The steps of the algorithm are given in detail. The proposed algorithm is applied to some classical Benchmark functions optimization and compared with the basic PSO(bPSO). Simulation results show IS-PSO can maintain better swarm diversity, avoid the premature convergence effectively and improve the global performance of PSO in solving the complex functions optimization.