Optimizing Task Assignment among Unmanned Vehicles in Terms of Energy Consumption

Abstract. This paper considers the topical issue of ensuring the availability of unmanned vehicles (agents) in a dynamic technical system (DTS) within an intelligent transport environment. The problem of uneven load distribution among agents, causing inefficient energy consumption and reducing the total operating time of the system, is studied. This problem is solved by proposing an optimization model that includes an objective function (maximizing the total operating time of the entire DTS) and a set of constraints describing the available energy of each agent. The key aspect of the model is to ensure the uniform distribution of the energy load among all agents. The optimization problem is solved using the CP-SAT Boolean Satisfiability algorithm with integer constraints. According to the experimental results with the CP-SAT algorithm, there is an interesting phenomenon, i.e., a correlation between the sampling step (the time interval during which the algorithm searches for an acceptable solution) and the execution time of the optimization program. Based on this correlation, a heuristic method for changing the sampling step is proposed. The study is primarily focused on the performance of the model and optimization algorithm in real conditions of robotic transport systems with exogenous disturbances. According to the testing results, the model demonstrates good performance on virtual agents with completely known system parameters and on a group of real agents (wheeled robots), where the system parameters are subject to disturbances.

Keywords: dynamic technical system, availability, optimization, distribution of energy resources, unmanned vehicle, safety.

Acknowledgments. This work was supported by the Russian Science Foundation, project no. 23-19-00338, https://rscf.ru/en/project/23-19-00338/.