Synthesis of Cost-optimized Controllers from Scenario-based GR(1) Specifications

Abstract

Modern systems often consist of many software-controlled components which must cooperate to fulfill difficult to achieve goals. Each component must properly react to both, external events, e.g., user inputs, and internal events, i.e., actions of other components. Making each component behave such that the cost of running the system is minimized, e.g., by minimizing total energy consumption, adds additional complexity. To support engineers in the difficult design of such systems we developed a scenario-based specification approach enabling the intuitive modeling of goals and assumptions using short scenarios. These specifications can be simulated and automatically analyzed, allowing defects to be detected and fixed early in the development process. In this paper we present and evaluate an extension to our approach which makes it possible to model costs of processes and thus to synthesize controllers which not only guarantee that the specified goals are fulfilled but are being fulfilled in a cost-optimized manner. Our approach even considers the transfer of energy between components to enable the design of systems in which, e.g., the braking energy of moving components can be leveraged to reduce the cost and environmental impact of a system instead of wasting this energy in the form of heat.