An investigation into the use of spatial coordinates for the genetic algorithm based solution of the fixture layout optimization problem

摘要

The fixture layout optimization problem is highly modal. It has been shown that layout optimization methods that use the genetic algorithm (GA) are effective at finding high quality solutions. To date, such methods have used FEA node numbers to represent the locations of fixture elements. This paper presents a new GA based optimization method that uses spatial coordinates to represent the locations of fixture elements. This method integrates a number of important GA concepts such as real encoding, increased mutation, and sharing. Test results show that the new method provides higher quality solutions with substantially less numbers of generations than the old method. These results also show that the use of increased mutation and sharing significantly improves the performance of the new method.