Power generation expansion planning with adaptive simulated annealing genetic algorithm

摘要

In this paper, an adaptive simulated annealing genetic algorithm is proposed to solve generation expansion planning of Turkey's power system. Least-cost planning is a challenging optimization problem due to its large-scale, long-term, nonlinear, and discrete nature of power generation unit size. Genetic algorithms have been successfully applied during the past decade, but they show some limitations in large-scale problems. In this study, simulated annealing is used instead of mutation operator to improve the genetic algorithm. The improved algorithm is applied to the power generation system with seven types of generating units and a 20-year planning horizon. The planning horizon is divided into four equal periods. The new algorithm provides approximately 6.6 billion USS (3.2 percent) cheaper solution than GA and also shows faster convergence.