(185o) Optimizing Energy System Design Using a Parallel Tabu Search Algorithm

The goal of this work is to develop novel parallel computing strategies for the solution of multi-objective energy production problems. The Tabu search algorithm provides a hierarchical approach, and the efficacy of the method derives from the ability of the algorithm to navigate to and escape from local minima. In this project, the Tabu search method is applied to Fazlollahi et al.’s [1] model for a multi-objective optimization problem involving energy design, which simplifies a potentially complex energy system by classifying energy conversion technologies into six subgroups which can then be optimized. The search employs the methods of Lin et al. [2] to minimize both environmental effects and cost, producing Pareto optimality curves which mimic those produced by the evolutionary algorithm originally utilized by the model’s author. This algorithm was implemented in the Python programming language. Many problems in terms of design and scheduling of energy systems can be formulated as MINLP’s, while objectives are formulated for cost and environmental impact. In this work, we apply results on both single and multiple processors to show the parallel efficiency of the approach.

References

[1] Fazlollahi, S., Mandel, P., Becker, G., Marechal, F. (2012). Methods for Multi-objective Investment and Operating Optimization of Complex Energy Systems. Energy, 45, 12-22.

[2] Lin, B., Chavali, S., Camarda, K., Miller, D.C. (2004). Computer-aided molecular design using Tabu search. Computers and Chemical Engineering, 29, 337-347.