论文部分内容阅读
把约束函数作为目标函数,将约束优化问题转化为多目标规划问题。对这个多目标规划,根据带权极小极大策略构造了一个同进化代数有关的变适应值函数。利用广义球面坐标变换和均匀设计法来选择权重,使得由此权重确定的适应值函数能使种群中的容许解逐渐增加并且保持其多样性。用均匀设计法构造的带有自适应性的变异算子增强了算法的局部搜索能力。该方法能有效处理约束,特别是紧约束。计算机仿真显示了该方法是有效的。
Using the constraint function as the objective function, the constrained optimization problem is transformed into a multi-objective programming problem. For this multi-objective programming, we construct a fitness function that is related to evolutionary algebra according to the strategy of minimum weight. The weights are selected by using the generalized spherical coordinate transformation and the uniform design method so that the fitness function determined by this weight can gradually increase the allowable solution in the population and maintain its diversity. The adaptive mutation operator constructed by the uniform design method enhances the local search ability of the algorithm. This method can effectively deal with constraints, especially tight constraints. Computer simulation shows that this method is effective.