Formal Verification of Ethical Choices in Industrial CPS

This paper addresses the issue of formal verification of ethical choices in Industrial Cyber-Physical Systems. An innovative approach based on Beliefs, Desires, and Intentions (BDI) agents and model checking is proposed. To do so, we first give a formal definition of ethical rules. Based on this definition, an algorithm is then designed to implement ethical reasoning. Finally, we apply this approach to TRACILOGIS (an academic full-sized application platform) to illustrate its feasibility. Four properties are designed and checked. The verification results show the agent with ethics can always reason out the least unethical actions to take.