Кey Words: Case-Base Reasoning (CBR); mathematical modeling; operational conditions; scheduling; suboptimal control; Thermal Treatment Process (TTP).
Abstract. An intelligent system for control of the thermal treatment process (TTP) of wood materials addressed toward manufacturing with necessity of often rescheduling is proposed via combination of model-based and data-driven approaches. Using First-principle mathematical model of TTP presented by Partial Differential Equations in 2D space with suboptimal model-based control algorithm and Case-Based Reasoning (CBR) approach an explicit suboptimal control system is investigated in different operational conditions. A set of virtual subspaces for feasible operational situations for variety of objective criteria of value assessment is created using traditional problem-decision representation. As the search spaces are well structured, the search procedure based on traditional K–NN algorithm is strongly simplified. In this way the complicated computer simulation of the TTP at each time step due to the plant’s parameter distribution, nonlinearity and operational or environmental disturbances are fulfilled off-line. On-line are accomplished relatively small part of the calculations connected with the traditional R4–operations in CBR, objective functions estimation, some databased and rule-based control parameter corrections and possible adaptation from charge to charge. Some results of the simulation experiments are presented and analyzed.