Hybrid Genetic Algorithm And Their Applications

Genetic algorithm is a random search method based on the mechanism of naturalselection and population genetics, which has broad applicability. Due to the weaker ability forlocal search and premature convergence of genetic algorithm, new ideas and methods areintroduced, a fast genetic algorithm and genetic-ant colony hybrid optimization method aredesign. The proposed algorithm are applied to solve TSP and the large automated warehousepicking path optimization.The main content of this thesis is as follows:(1) A fast genetic algorithm is designed. According to the characteristics of constraintand the engineering practice of large warehouse, the handling of constraints were distributedto the initial population generation and each genetic link. Single-point crossover is adoptedbetween the nodes of same location of different chromosomes. Combined with the constraintcondition and objective, a new partial-order relation is introduced for comparison ofindividuals. The elitism strategy is used to make the best individuals parent into offspring,thus, the algorithm has strong robustness.(2) A new genetic-ant colony hybrid optimization method is proposed based on fastgenetic algorithm. The serial structure is introduced into this algorithm, and the ant colonyalgorithm (ACA) is applied to generate the initial population, thus the optimal solution isfound depend on GA. Through employing the list of candidate cities in the initial pheromonematrix to decrease inferior solutions, and narrow the searching range of the solution space. Inorder to improve the local search ability of GA, the evolutionary reversal operation isintroduced into GA. The elitism strategy is adopted to avoid the degradation of offspring andimprove the diversity of population.(3) This hybrid genetic algorithm is used to solve TSP. The proposed method is verifiedto be valid by comparing result with the result of references.This fast GA is used to completethe optimal Pareto solution of optimization of the stacker picking path in an AutomatedStorage/Retrieve System (AS/RS) of a large airport, where there were some goods in storage.The simulation results indicated that the proposed algorithm reduced the computationalcomplexity of time and space greatly, found the optimal solution quickly and steadily. Whenthe iteration is stopped in the obtained Pareto solution set, the individual that meets the needsshows smaller (better) objective function value and dynamic adjustment with the change instorage locations, and meet the needs of practical engineering of AS/RSs optimization control.