Application of Swarm Intelligence Algorithms to Optimize the Power Consumption Model

This paper presents the problem of power consumption mathematical model development using the Pamir, the region of Tajikistan, as a case. The model view illustration describing the load curve is known, it is required to find the model parameters values. The paper compares three approaches: manual selection; deterministic method based on Fourier transform with local gradient search; meta-heuristic swarm algorithms. It is shown that swarm algorithms, due to their multipurposeness and scalability, make it possible to obtain more accurate models with less labor costs. But it is necessary to use several swarm algorithms, since it is impossible to determine in advance which one will be the best solution for a specific task. The results were confirmed by the application for the load curves of the UPS of Siberia.