Integrating genetic algorithms, tabu search, and simulatedannealing for the unit commitment problem

摘要

This paper presents a new algorithm based on integrating geneticnalgorithms, tabu search and simulated annealing methods to solve thenunit commitment problem. The core of the proposed algorithm is based onngenetic algorithms. Tabu search is used to generate new populationnmembers in the reproduction phase of the genetic algorithm. A simulatednannealing method is used to accelerate the convergence of the geneticnalgorithm by applying the simulated annealing test for all thenpopulation members. A new implementation of the genetic algorithm isnintroduced. The genetic algorithm solution is coded as a mix betweennbinary and decimal representation. The fitness function is constructednfrom the total operating cost of the generating units without penaltynterms. In the tabu search part of the proposed algorithm, a simplenshort-term memory procedure is used to counter the danger of entrapmentnat a local optimum, and the premature convergence of the geneticnalgorithm. A simple cooling schedule has been implemented to apply thensimulated annealing test in the algorithm. Numerical results showed thensuperiority of the solutions obtained compared to genetic algorithms,ntabu search and simulated annealing methods, and to two exact algorithms