Stepwise Comparison for Minimizing Controller Makespan
Yuki Shimizu, Takanori Hirano, Takuto Yamauchi, Kenji Tei · 2023
Controllers generated by controller synthesis are guaranteed to achieve a given goal in the system’s operating environment on the basis of game theory. However, while these controllers are guaranteed to be secure, they are not implemented to express preferences. Existing work implements a qualitative comparison framework that focuses on the makespan of discrete-event-based controllers concerning reachability goals. However, this framework focuses on the end-to-end process of the controller, so the computation time explodes with the size of the controller. Thus, we propose an algorithm that reduces computation time by dividing the controllers for which preferences are inferred into several processes and minimizing each makespan. This algorithm divides end-to-end processes into several processes by branches of controller and subgoals to the goal state. The stepwise comparison algorithm shows a time reduction of up to 91.1% over the models treated in minimizing makespan framework.