Forecasting Indoor Temperature Using Fuzzy Cognitive Maps with Structure Optimization Genetic Algorithm

Abstract

Fuzzy cognitive map (FCM) is a soft computing methodology that allows to describe the analyzed problem as a set of nodes (concepts) and connections (links) between them. In this paper the Structure Optimization Genetic Algorithm (SOGA) for FCMs learning is presented for prediction of indoor temperature. The proposed approach allows to automatically construct and optimize the FCMmodel on the basis of historical multivariate time series. The SOGAdefines a newlearning error function with an additional penalty for coping with the high complexity present in anFCMwith a large number of concepts and connections between them. The aim of this study is the analysis of usefulness of the Structure Optimization Genetic Algorithm for fuzzy cognitive maps learning on the example of forecasting the indoor temperature of a house. A comparative analysis of the SOGA with other well-known FCM learning algorithms (Real-Coded Genetic Algorithm and Multi-Step Gradient Method) was performed with the use of ISEMK (Intelligent Expert System based on Cognitive Maps) software tool. The obtained results show that the use of SOGA allows to significantly reduce the structure of the FCM model by selecting the most important concepts, connections between them and keeping a high forecasting accuracy.