论文部分内容阅读
现有遗传算法难以充分利用知识引导复杂的搜索寻优,无法有效求解挖掘机动臂结构优化问题。为提高动臂结构优化的求解效率与质量,提出一种基于知识的实数编码新型遗传算法。构建知识进化与遗传算法相集成的双重进化机制,实现优化过程浅层与深层隐性知识的挖掘、处理和利用,引导遗传算法的搜索寻优。所构建的双重进化机制通过知识利用算子实现知识进化与群体进化的有效结合,提高知识与遗传算子的可配置性。提出挖掘机动臂结构优化设计领域文化与优化过程知识集成引导的新型选择算子、交叉算子和变异算子,并以中型液压挖掘机动臂结构优化为例,通过八种知识引入程度不同的遗传算法进行对比验证。优化结果表明,将知识引入遗传算子,充分利用知识引导优化搜索可明显提高遗传算法的进化效率和搜索能力。所提出的基于知识引导的新型遗传算法将多层次知识进化与数值优化相结合,可为有效求解复杂工程优化问题提供新方法。
Existing genetic algorithms can not make full use of knowledge to guide complex search optimization, and can not effectively solve the structural optimization of excavator boom. In order to improve the efficiency and quality of boom structure optimization, a new genetic algorithm based on knowledge real number coding is proposed. The dual evolutionary mechanism of integrating knowledge evolution and genetic algorithm is constructed to realize the mining, processing and utilization of shallow and deep tacit knowledge in the optimization process and to guide the searching of genetic algorithms. The dual evolutionary mechanism constructed by this method can effectively combine knowledge evolution and group evolution through knowledge utilization operators and improve the configurability of knowledge and genetic operators. This paper presents a new selection operator, crossover operator and mutation operator guided by knowledge integration of culture and optimization process in the field of optimization design of boom structure of excavator. Taking the boom structure optimization of medium hydraulic excavator as an example, through the introduction of eight kinds of knowledge, Genetic algorithm to verify. The optimization results show that introducing knowledge into genetic operators and making full use of knowledge to guide optimization search can obviously improve the evolutionary efficiency and search ability of genetic algorithms. The proposed new genetic algorithm based on knowledge guidance combined with multi-level knowledge evolution and numerical optimization can provide a new method for solving complex engineering optimization problems effectively.