Abstract:

To solve global optimal solutions of nonlinear optimization problems, a new horizontal structure competition- mutually beneficial community optimization algorithm (HS-CBCO), is proposed based on the theory of horizontal structure competition-mutually beneficial community dynamics. In this algorithm, each biological population is composed of many biological individuals, the interaction across populations is mainly competition and mutual benefit.Within a population, there are interactions among individuals. Six operators are developed by using the community dynamics, among them, the competition and mutually beneficial operator can exchange information among individuals across populations, while the general and strong influence operator can exchange information among individuals within a population, thus realizing the full exchange of information among individuals. The newborn operator can timely supplement new individuals to a population, and the death operator can timely eliminate weak individuals from a population, thus greatly improving the ability of the algorithm to escape from local traps. The test results show that HS-CBCO has excellent exploitation ability, exploration ability and coordination between them, and has the characteristics of global convergence. The algorithm provides solutions for solving global optimal solutions of some complex function optimization problems.

Key words: swarm intelligence optimization algorithm, horizontal structure competition-mutually beneficial community dynamics, population dynamics, global optimal solution

摘要：

为了求解一些非线性优化问题的全局最优解，采用水平结构竞争-互利群落动力学理论，提出了一种新的水平结构竞争-互利群落优化算法（HS-CBCO）。在该算法中，每个种群由若干生物个体组成，种群间相互作用主要是竞争和互利，种群内部各个体之间存在相互影响。运用群落动力学理论开发出了6个算子，其中竞争和互利算子可实现个体跨种群交换信息，而普通影响和强烈影响算子可实现种群内的个体之间的信息交换，从而确保了个体间的信息的充分交换；新生算子可适时补充新个体到种群中，而死亡算子可将种群中的虚弱个体适时清除掉，从而提升了该算法跳出局部陷阱的能力。测试结果表明，HS-CBCO算法的求精能力、探索能力及其两者的协调性均优良，且具有全局收敛性的特点，为复杂优化问题全局最优解的求解提供了解决方案。

关键词: 群智能优化算法, 水平结构竞争-互利群落动力学, 种群动力学, 全局最优解