A New State Reduction Approach for Fuzzy Cognitive Map with Case Studies for Waste Management Systems

Abstract

The authors have investigated the sustainability of Integrated Waste Management Systems (IWMS). These systems were modeled by Fuzzy Cognitive Maps (FCM), which are known as adequate fuzzy-neural network type models for multi-component systems with a stable state. The FCM model was designed of thirty-three factors to describe the real world processes of IWMS in as much detailed and as much accurately as possible. Although, this detailed model meets the requirements of accuracy, the presentation and explanation of such a complex model is difficult due to its size.

While there is a general consensus in the literature about a very much simplified model of IWMSs, detailed investigation lead to the assumption that a much more complex model with considerably more factors (components) would more adequately simulate the rather complex real life behavior of the IWMS.

As the starting point we used the thirty-three component model based on the consensus of a workshop of experts coming from all areas of the IWMS (operation, regulation, management, etc.) and the set goal was to find the most accurate real model that could be obtained by analyzing and properly reducing this – very likely too much detailed, or atomized – model.

In this paper, a new state reduction approach with three different metrics is presented. The practical aspects of the results gained by these methods are evaluated.