A Benders Decomposition Method for a Multi-stage Stochastic Energy Market Equilibrium Problem

In this thesis, a stochastic energy market equilibrium model is developed and implemented in GAMS. The model involves multi objective optimization and is solved as a mixed complementarity problem. To provide an efficient solution strategy, a Benders Decomposition method tailored for the modelled energy market problem is studied and implemented. The scalability of the implemented decomposition algorithm is investigated for different versions of the energy market problem, and the results are compared to the alternative of direct solution in GAMS. An overall result is that the decomposition method succeeded in finding correct solutions, and proved to be the fastest solver option for the larger instances of the tested problems. In addition, the results did also facilitate a discussion of possibilities for further improvements in the efficiency of the algorithm