Genetic algorithm and tabu search hybrid algorithm to co-scheduling model of three Gorges-Gezhou dam

Abstract

This paper provides a mathematic model for Three Gorges-Gezhou dam co-scheduling problem, based on full analysis of Three Gorges-Gezhou dam's actual needs, to maximize the total throughput of Three Gorges-Gezhou dam, to maximize the utilization ratio of shiplock area and minimize the total navigation shiplock waiting time under multiple constraints. Due to some disadvantages of genetic algorithm, such as prematurity, weakness of climb. This paper combines genetic algorithm with tabu search to make the new hybrid algorithm have the two algorithms' advantages. It searches new solutions in the multiple individuals which is the characteristics of the genetic algorithm, while it has memory ability and efficient hill-climbing capability which is the characteristics of tabu search. Then we add the memory function into the crossover operator of the genetic algorithm, and use tabu search as the mutation operator. Real operation data from our project shows that the genetic algorithm and tabu search hybrid algorithm outperforms the traditional manual scheduling in that the less computational time is taken, the area utilization ratio of the five shiplocks is increased, the waiting time of high-prioritized ships is shorten, and a better balanced and alternating run-mode is provided for the three shiplocks in the Gezhou dam. © 2009 Springer Berlin Heidelberg.