基于自适应粒子群算法的动态优化

Dynamic optimization based on adaptive particle swarm optimization algorithm

自适应粒子群算法(AdaptiveParticle SwarmOptimization,APSO)是一种参数自适应的种群智能算法。该算法以种群的分布状态为依据区分优化过程中的不同状态自适应地调整算法参数。基于APSO算法具有参数自适应、快速收敛、全局搜索能力强等优点。将APSO算法应用于动态优化,通过采用按变量比例分配时间的方法构造时间变量,从而将其转化为无约束变量,通过时间变量与控制变量构造控制输入函数控制动态系统,使其达到最优。该方法提供一种转换时间变量约束的方法,使其能够作为一般优化问题,适用于其他类似演化类算法的动态性能的测试。最后,通过4个经典动态优化测试函数,比较APSO算法与蚁群算法,体现APSO算法处理动态优化的性能。

Adaptive particle swarm optimizatioin （APSO） is a swarm intelion algorithm, which regulates parameters adaptively during the optimization process according to the distribution state of the swarm particle. Due to its advantages of paramater adaptively regulating, fast converging and golbal optimizing, we apply the APSO algorithm into the dynamic optimization problems. We free the constraint of the time variable through the method of proportional dividing method, and contruct the control function combing input and time variables. Thanks to this method we change this kind of dynamic problems into a basic optimization problems, which makes them applicable to the test of other intelligence based optimization algorithms. In the end, we compare simulation result with the Ant colony algorithm which has the similar mechanism with PSO.