约束自适应桁架优化设计方法

摘要

A novel optimization algorithm by name Ω-CMA-ES, combining Oracle penalty function and metaheurastic algorithm was proposed to solve a typical multi-modal and highly non-linear problem,that is,the truss size and shape optimization.The algorithm can alleviate the cumbersome burden of parameters setting,and thus adaptively handle the constraints in truss design.Only one parameter Ω needs to be set manually while this algorithm is applied to handle various complex truss optimization problems.Numerical examples show the robustness of the proposed algorithm with respect to parameter Ω,for the algorithm can effectively handle various types of dynamic constrains,and the potential of the proposed algorithm in finding global optimal solution,for the relevant performance indicators are better than the results published in the literature.%为求解多峰值、高度非线性桁架尺寸及形状优化问题，减少算法参数设置的盲目性，将 Oracle 罚函数与启发式算法相结合，提出可自适应处理约束列式的优化算法Ω-CMA-ES。该算法在处理各类复杂桁架优化问题时仅需设置一个参数Ω。测试算例表明，该算法对参数Ω具有良好的鲁棒性，可有效处理各类动态约束；且在探索全局最优解时体现出较高潜力，优化质量及收敛速度优于既有结果。