求解离散优化问题的改进人工蜂群算法的研究

Improved Artificial Bee Colony Algorithm for Discrete Optimization Problems

目的 解决人工蜂群算法在求解连续优化问题时易陷入局部最优,收敛速度慢,而且算法在最初设计时不适用于离散问题的求解等问题。方法 而0-1问题属于典型的二进制离散约束优化的NP-Hard问题,故提出了基于离散优化问题的人工蜂群算法(DABC)。首先,采用二进制编码方法,改进解的编码形式;其次,使用多维邻域搜索策略改进ABC算法的搜索策略,并在雇佣蜂阶段引入高斯变异,保持种群的多样性,加快算法的收敛速度。在侦察蜂阶段引入柯西变异算子,以增强算法的全局搜索能力,避免算法在迭代时陷入局部最优,进一步提高算法的效率和精准度。结果 通过实验仿真验证了算法的有效性和高效性,当种群规模增大时,算法的收敛速度加快,从而验证了不同的参数值对算法的影响。结论 改进后的算法在求解离散优化问题时确保种群的多样性,提高了算法的收敛速度、整体寻优能力和开发能力。

Objective The artificial bee colony algorithm is mainly applied to solve continuous optimization problems,and it is easy to fall into the disadvantage of local optimum and slow convergence speed.Methods 0-1 is a typical binary NP-Hard problem for discrete constrained optimization.An artificial bee colony algorithm(DABC)based on discrete optimization problem was proposed.Firstly,the coding form of the solution was improved and binary coding was adopted.Secondly,the search strategy of the ABC algorithm was improved by using the multi-dimensional neighborhood search strategy,and the diversity of the population was maintained by introducing the Gauss mutation in the hiring bee stage.The Cauchy mutation operator was introduced to the reconnaissance bee stage to enhance the global searching ability of the algorithm and avoid to fall into the local optimal solution of the algorithm.The efficiency of the algorithm was further improved.Results The effectiveness and efficiency of the algorithm were verified by experimental simulations.When the population size increases,the convergence speed of the algorithm is higher,and the effect of different parameter values on the algorithm is verified.Conclusion The improved algorithm ensures the diversity of the population in solving the discrete optimization problem,and improves the convergence speed,the overall optimization ability and the development ability of the algorithm.