The Research Of Improved Geneticalgorithm For Plant Layout Optimization Problem

In the scientific research and engineering practice, many problems aredemonstrated as nonlinear programming problems. Traditional optimizationalgorithm can not solve these problems very well because of theirMulti-parameter, High complexity, uncertainty and other characteristics. Alsothe calculations are always very time-consuming and not satisfy the people'sdemand. As one of the nonlinear programming problem, plant layout problemalso has high dimensions, large constraints and other characteristics. In thispaper, the genetic algorithm are improved and used to solve this problem,which obtained a satisfactory solution.Genetic algorithm is a novel stochastic search algorithm which has a drawon Evolutionary Theory and Genetics ideas, and has a high Solution efficiencyand globe search ability. In this paper, first we reduced the number of variablesby simplified the non-overlapping and inclusion constraints, for the number ofvariables of the plant layout problems is usually large. Second, on theconstraints handling problem, a Constraints scattered method is proposed toreduce the difficulty of generate feasible individuals subject to manyconstraints. Meanwhile, the individual evaluation criterion is improved and a feasible-infeasible crossover method is proposed to enhance the Improvesolution accuracy and convergence rate. At last, the steepest descent method isintroduced as a means to strengthen the local search. The experiments showthat our improved genetic algorithm has an outstanding solution accuracy andefficiency, which is better than other three state-of-the-art approaches. In thecase study of plant layout problem, the result of our method can satisfy thenon-overlapping, minimum safety distance and toxic gas leakage constraints,and give an ideal solution.