Using nonlinear constrained optimization methods to solve manipulators path planning with hybrid genetic algorithms
Xinglong Zhu, Hongguang Wang, Mingyang Zhao · 2005
Numerical optimization problems enjoy a significant popularity in genetic algorithms (GAs) community. All major genetic techniques use such problems for various tests and experiments. However, many of these techniques encounter difficulties in solving some real-world problems which include non-trivial constrains. This paper discusses a new method, which combines sequential weight increasing factor technique (SWIFT) with GAs, for solving nonlinear constrained optimization problems. In order to surmount the pre-maturity phenomenon, the niche evolutionary strategy is adopted. By comparison of individuals in the same generation computation, if the individual is fit for the differentiate criterion, the lower fitness individual will decrease its fitness value on use of penalty methods. Eventually, some famous test cases and manipulators planning illustrate this approach is very available