Modelling physical systems using fuzzy inference cognitive maps

When faced with real world problems information is often limited and comes in various forms. Information about a system may come in the form of expert opinion, theoretical analysis or empirical data. These problems can often contain cyclic interactions. In this paper a novel approach is proposed based around a development of the fuzzy cognitive map (FCM). The cognitive map is produced according to expert knowledge. This allows an intuitive and structured representation of the problem to be made. This can include representation of interactions and feedback. Adaptive neuro fuzzy inference systems (ANFIS) are used to model the relationships. The use of ANFIS means both qualitative (expert knowledge) and quantitative (empiric data) sources can be used to define the causal relationships. This technique is referred to as a fuzzy inference cognitive map (FICM). High efficiency deep grinding is a new grinding regime. Only limited empirical data is available. Initial theoretical models have been developed they include a circular interaction. This paper outlines how HEDG can be modelled using FICM. This incorporates analytical models, expert opinion and experimental data into one model. The paper presents initial results.