多策略蜂群算法及在数字系统建模中的应用

Multi-strategy Based Artificial Bee Colony Algorithm for Digital System Modeling

为提高人工蜂群算法的全局搜索能力,提出一种多策略蜂群算法(multi-strategy based artificial bee colony algorithm:m ABC)。该算法设计一种基于最优蜜源的位置计算策略,然后与经典蜜源位置计算策略合作,在引领蜂阶段,以随机进化模式对蜜源进行更新,在跟随蜂阶段,以组合进化模式对蜜源进行更新。同时,设计一种基于差分进化算子的侦察蜂进化模式。对10个经典测试函数和2个数字系统建模问题进行实验仿真,实验结果表明,相比标准蜂群算法,m ABC算法模式有能效地平衡算法的探索和开发能力,提高收敛速度和最优解的精度,具有良好全局搜索效率,是一种有效的求解全局优化问题的方法。

To improve the search ability of artificial bee colony algorithm （ABC）, this paper proposes amulti-strategy ABC named mABC here. The new mABC includes a best solution based position calculationequation and then during the employed bees phase, the best solution based strategy and the conventionalstrategy work together in a random model and during the onlooker bees＇ phase, the two strategies worktogether in the composite model. Furthermore, a new differential evolution inspired model is designed inscouts phase. The simulation on 10 benchmark functions and 2 digital system modeling problems shows that,compared with the conventional ABC, the proposed framework of mABC effectively balances theexploitation and exploration ability, and improves the accuracy of the optimal solution and the convergencespeed, mABC has remarkable global search capability, and is a potential method for solving globaloptimization problems.