Hybrid Filtered Beam Search Algorithm for the Optimization of Monitoring Patrols

This paper offers an operational and methodological response for managing industrial risks by improving the monitoring of industrial areas. The objective is the optimization of monitoring patrols with automated mobile agents that are responsible for the surveillance. Such agents are formed by automated guided vehicles or unmanned aerial vehicles that carry various sensors. Apart from the specificities of each class of agents, the proposed approach is motivated by the need to inspect sites that may be dangerous or difficult to access. The optimization of the missions is carried out in compliance with functional (e.g., precedence of the operations) and operational (e.g., the travel time reserve of the agents) constraints in the double perspective of patrol configuration and trajectory planning as far as these aspects are strongly correlated. The questions that should be answered are as follows. How many mobile agents are required to perform a given set of measurements? How many sensors and what types of sensors must each of these agents equip? How to define the mission and trajectory of each agent? Such questions are studied as a multi-robots / multi-tasks problem, and an approach based on the hybrid filtered beam search is proposed for that purpose.