INTELLIGENT SYSTEMS IN INDUSTRIAL PRODUCTION PROCESS SCHEDULING Agnieszka Wielgus ∗ Rados law Rudek ∗ Iwona Po´ zniak-Kosza lka ∗∗ Leszek Kosza lka ∗∗ Keith J. Burnham ∗∗∗ ∗ Institute of Computer Engineering, Control and Robotics, Wroclaw University of Technology, Janiszewskiego 11/17, 50-372 Wroclaw, Poland ∗∗ Chair of Systems and Computer Networks, Wroclaw University of Technology, Wyb. Wyspianskiego 27, 50-370 Wroclaw, Poland ∗∗∗ Control Theory and Applications Centre, Coventry University, CV1 5FB, UK Abstract: This paper is devoted to real life flexible industrial production systems using universal machines, which are able to perform different kinds of jobs. In such environments, jobs from various families require different sets of processing facilities, thus, some rearrangements are needed in case of switching between families. To solve this problem, we formulate this issue in the context of scheduling theory. However, the problem is NP-hard, therefore, it is highly unlikely to find an optimal solution in polynomial time, hence, we provide some heuristic algorithms and one based on a simulated annealing approach, together with computational verification of their effectiveness. Keywords: Industrial production systems, Scheduling, Heuristics 1. INTRODUCTION Modern production and manufacturing systems are very often based on universal machines, which are able to perform different kinds of jobs. In such systems, jobs, which require the same sets of production facilities belong to one family. How- ever, a rearranging (setup) of a machine is needed whenever there occurs a switching between jobs belonging to different families, which on the other hand is related with additional time and cost of production. Although, such production systems are flexible, the optimization of a production line becomes more difficult than in classical systems, with fixed, dedicated machines. These practical industrial problems have attracted significant interest and extensive research in this domain has been carried out (e.g., (Potts and Kovalyov, 2000), (Ng et al., 2005)). Whilst, these problems are extensively studied in the scientific literature form the theoretical point of view, in this paper, we focus on on a real life industrial problem, providing algorithms, which operate on real data. The remainder of this paper is organized as fol- lows. Section 2 contains the problem formulation in the context of scheduling theory. A description of the algorithms appears in Section 3 and their