Control of discrete-event systems with modular or distributed structure

Most of the large scale state transition (also called discrete-event) systems are formed as parallel compositions of many small subsystems (modules). Control of modular and distributed discrete-event systems appears as an approach to handle computational complexity of synthesizing supervisory controllers for large scale systems. For both modular and distributed discrete-event systems sufficient and necessary conditions are derived for modular control synthesis to equal global control synthesis, while enforcing a safety specification in an optimal way (the language of the controlled system is required to be the supremal one achievable by an admissible controller and included in a safety specification language). The two cases of local (decomposable) and global (indecomposable) specifications are considered. The modular control synthesis has a much lower computational complexity than the corresponding global control synthesis for the respective sublanguages. The complexity is compared using explicit formulas.