论文部分内容阅读
蚁群算法作为一种新的智能计算模式,由于其离散性本质而在组合优化问题上取得巨大成功,但这也限制了它在连续问题求解中的应用。为此,提出一种用于连续域寻优的改进蚁群算法。算法的局部搜索基于解决离散问题的经典蚁群优化思想,全局搜索利用AntWalk和AntDiffusion技术,且每代寻优结束后均采用“精英策略”把本代最优个体保留到下一代中。最后在理论上对其进行了收敛性分析,证明可较快地收敛到全局最优解,并用几个基准函数对算法做了仿真测试,均取得良好效果。
As a new intelligent computing model, ant colony algorithm (ACO) has great success in combinatorial optimization because of its discrete nature, but it also limits its application in solving continuous problems. To this end, an improved ant colony algorithm for continuous domain optimization is proposed. The local search of the algorithm is based on classical ant colony optimization which solves the discrete problem. The global search uses AntWalk and AntDiffusion techniques. After the optimization of each generation, the “elite strategy” is adopted to save the best individual in the generation to the next generation. Finally, the convergence analysis of the proposed method is carried out in theory, which proves that it can converge to the global optimal solution quickly and simulates the algorithm with several benchmark functions, all of which have achieved good results.