Evolutionary Multi-Objective Algorithm to effectively improve the performance of the classic tuning of fuzzy logic controllers for a heating, ventilating and Air Conditioning system

In this work, we present an advanced Multi-Objective Genetic Algorithm for obtaining more compact fuzzy logic controllers as the way to find the best combination of rules, thus improving the system performance in a problem to control a Heating, Ventilating, and Air Conditioning system. To this end, two objectives have been considered, maximizing the performance of the system (involving energy performance, stability and indoor comfort requirements) and minimizing the number of rules obtained (for finding the most cooperative/accurate rule subset). This algorithm is based on the well-known SPEA2 but implements some advanced concepts such as incest prevention which, by avoiding unnecessary evaluations, helps to improve the exploration/exploitation trade-off and consequently its con vergence ability, and makes use of different mechanisms to focus the search on the desired Pareto zone. All of this becomes essential in the problem addressed due to the long computation time required for each controller evaluation. The results obtained confirm the effectiveness of our approach in terms of the controller performance when it is compared with several mono-objective steady-state genetic algorithms, previously applied to this particular problem, and with the corresponding multi-objective counterparts, obtaining the most accurate fuzzy controllers by means of searching for simpler solutions.