Multistage fuzzy control using trapezoidal membership functions and dynamic programming

This paper presents a practical approach for the control error calculation under fuzziness as an extension of fuzzy decision making regarding aspects of robustness and applying multistage fuzzy control. Instead of fuzzy goals and constraints, stationary and dynamical fuzzy goals are used for solving dynamic programming. All presented concepts are based on operations using trapezoidal membership functions; therefore an obvious method for weighting of fuzzy goals is presented as well as a method to force a solution in tasks with contrasting requirements. The concept is applied in simulation case studies for optimization of temperature setpoints of heating systems in the field of preventive conservation, where special climate requirements are present.