On Value-Aligned Cooperative Multi-agent Task Allocation

Ensuring trustworthy AI and adopting value-aware systems with ethical principles are essential for the harmonious coexistence of humans and intelligent systems. This paper proposes a value-aligned cooperative multi-agent task allocation model for a cooperative whose members jointly own and operate the organization, working towards common goals. The proposed approach instantiates human values into ethical and moral norms. We focus on value awareness through introducing ethical values like justice and fairness into a mathematical programming model for near-optimal multi-agent task allocation. Key performance indicators are modeled using diverse measures in the task sharing context. A simple use-case example demonstrates how value awareness integrates into a decision support system for multi-agent task allocation in a cooperative agri-robot fleet context.