A Parallelized Genetic Algorithms approach to Community Energy Systems Planning

Optimization is being applied to a large number of disciplines and fields when there is an interest in finding the best solution among all the feasible solutions. It can be implemented through several methods. Genetic Algorithms are a case in point. They are mainly used for the energy resources scheduling problems as they help selecting the best schedule using a fitness function. Genetic algorithms are very useful for optimization problems with thousands of potential solutions, but, at the same time, their execution usually takes a long time. Hence, there is a need to look into potential performance improvements. In this paper, we are presenting the use of Genetic Algorithms as an optimization technique in a research project that performs the economic feasibility analysis for the integration of renewable energies into different types of buildings as part of the energy community system's planning. The outcome of the project is meant for non-technical users with limited technical knowledge who will be submitting their requests and expecting an output in a relatively short period of time. The performance of the tool is crucial for a satisfactory user experience. In order to overcome this issue, the work presented in this paper investigates the use of parallelization and distributed computing to decrease the overall response time. This process is double-fold: it requires making the code parallelizable and then running it in a cluster of distributed computers. In this work, we present the results of deploying the algorithms without parallelization and compare them to the results obtained when applying the parallelization. The results show that the parallelization has a great impact on the response time as it significantly drops.