Ontological model of work planning processes at manufacturing enterprises within a multi-agent decision support system

This paper presents an ontological model of work planning processes at manufacturing enterprises, integrated into a multi-agent decision support system based on a modular approach. The multi-agent system utilizes a coalition-holonic architecture representing a multi-level structure of agents grouped into static holons, managed by supervisors for recurring tasks and temporary coalitions, managed by coordinating agents for solving complex tasks. The ontology serves as a shared semantic space and a knowledge source for agents, ensuring cognitive coherence of their actions and serving as a means of coordinating holons and coalitions. Interaction with the ontology is implemented through specialized agents. The adjustment of decision criteria priorities is performed via weighting coefficients, which enhances the adaptability and controllability of the system. The cognitive load on decision makers is reduced through the ontological elimination of inconsistent alternatives and predefined rules. These effects are particularly pronounced with increasing resource scarcity and stricter constraints imposed on the generated plan. The results of the system's pilot testing suggest potential reductions in planning time, a more even distribution of workload among production personnel, and improved compliance with budget constraints during work planning. The practical significance of the proposed approach lies in accelerating the preparation and increasing the accuracy of production work plans, and reducing the cognitive load on decision makers when planning work for employees of small and medium-sized manufacturing enterprises.