一种考虑环境作用的协同免疫遗传算法

在综合考虑了环境对生物进化的影响、免疫算法的结构以及遗传算法部分算子的基础上,提出一种考虑环境作用的协同免疫遗传算法(ESIGA),以实现提高算法搜索速度和全局搜索能力的目标.在该算法中,设计了克隆环境演化算子和自适应探索算子,并构造了3个子种群协同进化以发挥克隆环境演化算子的影响,从而提高算法的全局搜索能力.引入的自适应探索算子和克隆环境演化算子,使算法具备了一定的学习能力,可加速搜索和防止早熟.构建的主种群和协同种群相互影响,使得算法对环境具有改良能力,加强了克隆环境演化算子的性能,而精英种群则加强了算法在优质个体邻域的搜索能力.采用13个常用无约束优化问题测试函数对算法做了检验,测试数据表明:ESIGA算法与正交遗传算法相比,其搜索速度要快于正交遗传算法1～2倍,并能够处理1 000维的高维优化问题.

Environment-Based Synergic Immune Genetic Algorithm

The function of environment effecting creature evolution, the process of immune clone algorithm and some operators of genetic algorithm are considered synthetically. An environment-based synergic immune genetic algorithm (ESIGA) is presented to improve the searching rate and global searching ability of optimal algorithm. The clone-environment operator and self adaptive search operator are designed to make the algorithm possess learning capability to accelerate the searching rate and avoid prematurity. And three groups of sub-population are used for improving the global searching ability, where the interaction between main sub-population and synergie sub-population improves the environment and exerts clone-environment operator effect, and the elitist sub-population enhances the neighborhood searching. Then 13 classical benchmarks of non-con-straint optimal problem are executed by the proposed algorithm. The test results demonstrate that ESIGA searching rate is twice as QGA/Q. ESIGA is suit to deal with 1 000-dimension non-constraint optimal problems.