A performance evaluation algorithm for discrete event systems under blocking

Abstract

When the supervised language includes too many possible blockings, selecting the minimal restrictive nonblocking solution(MRNBS), which takes away all possible blockings, sometimes corresponds a conservative behavior. Then relaxing the nonblocking condition becomes an inevitable fact to improve the overall performance. At the other hand, selecting the complete satisfying solution (CSS) as supervisor may let serious system failures due to some possible blockings. Then a blocking supervisor between these two solutions has to be selected. So an equilibrium point has to be searched between blocking and achievement in a numerical manner. For this purpose in this work a new formulation, which includes a new performance measure and search algorithm, is proposed. The offered new algorithm explores the best result according to introduced performance measure.