Teaching–Learning‑Based Optimization Algorithm for Path Planning and Task Allocation in Multi‑robot Plant Inspection System

摘要

A proper planning of path along with task management makes a multi-robot system quicker in task completion and fuel efficient. However, the previous related literature has explored a little bit in optimization of completion time and fuel consumption simultaneously. In this study, a new A* algorithm integrated multi-objective Teaching–learning-based optimization (MOTLBO) algorithm is presented for path and task management of multiple robots in plant inspection system. Two objectives namely total completion time (TCT) and total fuel consumption (TFC) are considered for minimization in this study. The developed technique implements A* algorithm for robot route planning and TLBO algorithm for task allocation to the robots. Two mutation operators namely insertion operator and interchange operator are applied to update solutions in teacher phase and learner phase of discrete TLBO. An approach for sequencing of tasks allocated to the robots and a scheme to make the robots’ movement collision free is also integrated with the proposed algorithm. For investigating the suitability, an instance of tank firm inspection is solved using this algorithm. Later on, the efficiency of the algorithm is tested by comparing it with the two existing multi-objective optimization algorithms namely Non-dominated Sorting Genetic Algorithm Ⅱ (NSGA-Ⅱ) and Heuristic Coupled Particle Swarm Optimization (HPSO) algorithm. Comparison has also been made between two variants of single objective genetic algorithm with A* algorithm and the proposed technique. The results reveal that the proposed algorithm outperforms the existing algorithms.