A comparative analysis of various optimization models for electricity capacity expansion

Several optimization models have been proposed in the literature to solve the project selection, timing and sequencing problem. Models based on dynamic programming (DP) such as embedded state DP, objective space DP, and others have been used to solve the capacity expansion problem. Also, models based on mixed integer linear programming using Benders decomposition theory, and heuristic rules have been proposed to solve that problem. In this paper these models are analyzed and compared from the theoretical point of view. All the reviewed models are then compared using them to define the minimum cost expansion of the Colombian electric sector. The Colombian interconnected electric generation system has about 10,000 MW of installed capacity with a generation composition of 80% hydro-generation plants and 20% thermo-generation plants. The expansion planning of the system has been carried out at a national level using simulation and optimization techniques. All these techniques are reviewed and discussed in this paper. Some conclusions and recommendations are finally presented.