Synthesis of self-adaptive supervisors of multi-task real-time object-oriented systems using developmental genetic programming

Abstract

This chapter presents a procedure for automatic creation of self-adaptive artificial supervisors of multi-task real-time object-oriented systems (MT RT OOS). The procedure is based on developmental genetic programming. Early UML diagrams describing a MT RT OOS are used as input data to the procedure. Next, an artificial supervisor which optimizes the system use is automatically generated. The supervisor is self-adaptive what means that it is capable of keeping optimality of the system in spite of disruptions that may occur dynamically in time of the systemwork. A representative example of creation of a supervisor of building a house illustrates the procedure. Efficiency of the procedure from the point of view of self-adaptivity of the supervisor is investigated.