Development of an Improved Genetic Algorithm and its Application in theOptimal Design of Ship Nuclear Power System

摘要

This article focuses on the development of animproved genetic algorithm and its application in theoptimal design of the ship nuclear reactor system, whosegoal is to find a combination of system parameter valuesthat minimize the mass or volume of the system given thepower capacity requirement and safety criteria. Animproved genetic algorithm (IGA) was developed using an'average fitness value' grouping + 'specified survivalprobability' rank selection method and a'separate-recombine' duplication operator. Combining witha simulated annealing algorithm (SAA) that continues thelocal search after the IGA reaches a satisfactory point, thealgorithm gave satisfactory optimization results from bothsearch efficiency and accuracy perspectives. This IGA-SAAalgorithm successfully solved the design optimizationproblem of ship nuclear power system. It is an advancedand efficient methodology that can be applied to the similaroptimization problems in other areas.