Application of multiobjective optimization and multivariate analysis in multiple energy systems: A case study of CGAM

Summary The performances of multiple energy systems are heavily dependent on its parameters. Selecting appropriate parameters and understanding the intrinsic relationships among parameters are important tasks for the designer. The study of the optimal parameter design of a benchmark case for multiple energy systems, a well‐known CGAM problem is presented. The integrated analysis process proposed here is composed of multiobjective optimization, decision making, and multivariate analysis. The conflicting performances, namely, energy, economy, and environment effects are considered as objectives simultaneously. An efficient multiobjective evolutionary algorithm, Adaptive Range Multi‐Objective Genetic Algorithm (ARMOGA) is introduced for finding the Pareto set. This method is based on a real‐coded multiobjective evolutionary algorithm, where the new design search range can be adjusted according to the statistics of former solutions. The range adaptation can help to reduce the number of function calls. The final compromise solution in the Pareto set can be selected by nearest to the utopian solution method. The characteristics of the Pareto solutions are investigated using multivariate analysis techniques. Clustering and dimensionality reduction approaches are employed for mining meaningful information between design variables and objectives. The numerical results demonstrate the effectiveness of the proposed analysis flowchart. These approaches can be extended to further application in real complex multiple energy systems.