Optimization of Multi-Modal Discrete Functions Using Genetic Algorithms
Duc Truong Pham, Yimin Yang · Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering · 1993
Four techniques are described which can help a genetic algorithm to locate multiple approximate solutions to a multi-modal optimization problem. These techniques are: fitness sharing, ‘eliminating’ identical solutions, ‘removing’ acceptable solutions from the reproduction cycle and applying heuristics to improve sub-standard solutions. Essentially, all of these techniques operate by encouraging genetic variety in the potential solution set. The preliminary design of a gearbox is presented as an example to illustrate the effectiveness of the proposed techniques.