A hybrid genetic algorithm/dynamic programming approach to optimal long-term generation expansion planning

Abstract

This paper describes an application of a hybrid approach to optimal generation expansion planning (GEP) using a refined genetic algorithm (GA) and the tunnel-based dynamic programming (DP). Long-term GEP is concerned with a highly constrained non-linear discrete dynamic optimization problem that can only be fully solved by complete enumeration, a process which is computationally infeasible in a real-world GEP problem. For this reason, commercial packages have searched a reduced solution space or applied mathematical programming algorithms, which result in being stuck in a local optimum. This paper proposes a hybrid approach combining a refined GA with the tunnel-based DP, a method employed in the Wien Automatic System Planning Package (WASP). The main advantage of this approach lies in the GA's capability to find the global optimum and the tunnel-based DP's high performance to get a local optimum. The framework developed can simultaneously overcome the “curse of dimensionality” and a local optimal trap inherent in the conventional mathematical programming approaches. The suggested method has been successfully applied to two test systems with 15 existing power plants, five types of candidate plant and a 14-year planning period, and a practical long-term system with a 24-year planning period.