Differential MAP-Elites Considering Energy Resource Management Behaviors in Power Distribution Networks

Problems like energy resource management take into account high dimensionality, nonlinear constraints, parameter uncertainty, or the dispersed nature of resources, and classical optimization tools that were not built for such a new paradigm are nearing the end of their useful life. In the process of striking a balance between “optimal” and “practical” answers, metaheuristic optimization emerges. This optimization presents some disadvantages as it does not guarantee the global optimum and lacks explainability in its processes, as it is considered a black-box algorithm. This work proposes a differential Multi-dimensional Archive of Phenotypic Elites (MAP-Elites) algorithm based on multiple differential evolution strategies to illuminate metaheuristic performance. We compare the multiple strategies for day-ahead energy resource management in a 33-bus distribution network with a high penetration of renewables and electric vehicle users. Results show that incorporating the differential MAP-Elites algorithm introduces an understanding of algorithm performance for the chosen behaviors.