Abstract: A new population - based search algorithm, Line-up Competition Algorithm(LCA) , is presented, which is found to be very efficient in solving non - convex nonlinear programming (NLP) and mixed-integer nonlinear programming (MINLP) problems. In LCA algorithm, there always exist independent and parallel evolutionary families. By two different competition fashions which are the competition for existence inside family and the position competition between the families, population is made to evolve rapidly toward optimal or near optimal region. The concept of competition driving force is proposed and its quantitative description is given. These families are arranged a line - up in the light of the value of their objective function, and allocated the corresponding search space according to their position in the line-up, resulting in the balance of local search and global search. The technique of contracting search space is applied for increasing the convergence speed. The algorithm was tested with a set of typical non - convex NLP and MINLP problems, the results obtained indicate that LCA has obvious advantage over genetic algorithms and simulated annealing algorithms in solution quality and search rate.

摘要： 提出了一种称为列队竞争算法(LCA)的群体搜索算法,该算法在进化过程中始终保持着独立并行进化的家族,通过家族内部的生存竞争和家族间的地位竞争这两种不同的竞争方式,使群体快速进化到最优或接近最优的区域.根据家族的目标函数值大小排列成一个列队,按列队中家族地位的不同分配不同的搜索空间,使局部搜索与全局搜索达到均衡,同时,应用逐步收缩搜索空间技术加速收敛速度.数值计算表明,列队竞争算法的搜索效率优于遗传算法和模拟退火法等算法.