最坏情况下的鲁棒优化问题研究

摘要

由于实际中存在噪声等不确定干扰，解的实际性能会受到很大影响，此时解的鲁棒性决定了解的实用性，鲁棒优化问题日渐成为国内外学者研究的一个热点。鲁棒优化问题与传统的最优化问题有着明显的区别，其考虑了实际应用环境往往不稳定且容易受到噪声等不确定因素的干扰。主要研究的是在不确定因素干扰下，采用协同演化算法寻找最坏情况下的鲁棒最优解。通过测试函数的实验验证了该方法思想的有效性。%Due to the existence of uncertain noise interference, the practical performance of the solution would be seriously influenced. However, the robustness of the solution determines the practicability of the solution, the robust optimization problem thus becomes the hot research topic both an home and abroad. The robust optimization problem is obviously different from the traditional optimization problem, for the actual environments are often unstable and susceptible to noises in real application. A novel method with coevolutionary algorithm to find the robust optimal solutions is proposed. The experiments on the test problem indicate that the proposed method is feasible and efficient.