Particle swarm optimization algorithm for the optimization of rescue task allocation with uncertain time constraints

This paper focuses on the problem of robot rescue task allocation, in which multiple robots and a global optimal algorithm are employed to plan the rescue task allocation. Accordingly, a modified particle swarm optimization (PSO) algorithm, referred to as task allocation PSO (TAPSO), is proposed. Candidate assignment solutions are represented as particles and evolved using an evolutionary process. The proposed TAPSO method is characterized by a flexible assignment decoding scheme to avoid the generation of unfeasible assignments. The maximum number of successful tasks (survivors) is considered as the fitness evaluation criterion under a scenario where the survivors' survival time is uncertain. To improve the solution, a global best solution update strategy, which updates the global best solution depends on different phases so as to balance the exploration and exploitation, is proposed. TAPSO is tested on different scenarios and compared with other counterpart algorithms to verify its efficiency.

Automated summary

This paper proposes a modified particle swarm optimization algorithm for robot rescue task allocation, which features a flexible assignment decoding scheme and uses the maximum number of successful tasks as the fitness evaluation criterion. A global best solution update strategy is also introduced to balance exploration and exploitation.