An improved heuristic algorithm FEIDEQ for the maximum legal firing sequence problem of Petri nets

The paper proposes a heuristic algorithm FEIDEQ for the maximum legal firing sequence problem of Petri nets (MAX LFS for short) and evaluates it based on experimental results. FEIDEQ is improved from FSDC that has been known to have the highest capability among the existing ones, by incorporating a new selection rule of transitions. In our experimental evaluation, FEIDEQ is applied to 2540 test problems, and, for each of 2040 of them, existence of an optimum solution is guaranteed. Among these 2040 test problems, FEIDEQ has produced an optimum solution to each of 1864 (91.3%) test problems, showing about 2.57 times that of FSDC, the best one so far. Furthermore, by adopting improved backtracking, average CPU time is about 0.86 times that of FSDC for general nets.