粒子动力学演化算法及其在求解SOP上的应用

A Particle Dynamical Evolutionary Algorithm and Its Application on Solving Single-Object Problems

根据粒子输运理论中粒子相空间能量最小原理和熵增法则提出一类粒子动力学演化算法,并在数据实验中用该算法求解了一些传统演化算法难于求解的单目标优化问题(SOP),比如求解凸函数和多峰函数等最小优化问题。由于这种新的动力学演化算法能使几乎所有的个体参加杂交和变异,因此,与传统的演化算法相比,该方法不仅能较快速地求解线性和非线性优化问题,而且还能容易地求出达到最优解的多个坐标点。

A particle dynamic evolutionary algorithm is presented based on particle transportation theory according to the principle of energy minimization and the law of entropy increasing in phase space of particles, and in the numerical experiments, we have used this algorithm to solve the single-object optimization problem (SOP), which the traditional evolutionary algorithm is difficult to solve. Compared with the traditional evolutionary algorithm, this algorithm not only can solve linear and non-linear optimization problems more quickly, but also can more easily solve the all coordinates reached the global solution of problems.