A Genetic Algorithm to Solve Day-Ahead Energy Market Clearing Problem Under Network Constraints

In this study, we work on the Turkish day-ahead electricity market (T-DAM) clearing problem. Similar to the other European markets, T-DAM is designed as a two-sided blind auction in which both buyers and sellers can submit their bids for different periods of the delivery day. The market operator has to decide on the accepted quantities of each bid, bid matchings, so that daily market surplus is maximized under periodic supply-demand balance constraints. However, the current T-DAM clearing problem does not take into account the physical transmission constraints. This situation poses serious problems in terms of supply and network security, and requires more costly congestion management in the real time. In this study, we extend the current T-DAM clearing problem by integrating transmission network constraints and more sophisticated bid types like profile block b ids and multi-period fl exible bi ds with time windows. We developed a genetic algorithm based math-heuristic to solve the extended T-DAM clearing problem. We tested our algorithm on specially generated random instances and compared with the performance of the state-of-the-art solver. The computational results show that the devised genetic algorithm is capable of finding an optimal solution in at least the third of the all test instances. It performs better than CPLEX MIQP solver on the hard-to-solve instances. The use of our genetic algorithm to solve T-DAM clearing problem can help to increase market surplus and thus promote market efficiency.