Optimal path planning of multi-robot in dynamic environment using hybridization of meta-heuristic algorithm

This paper investigates an innovative strategy for generating a collision and deadlock-free optimal position for the individual robots by satisfying the constraints of a dynamic environment for path planning of multiple mobile robots. The current study emphasized the shortfalls of the previous investigation on multi-robot path planning and offered an energetic methodology through the hybridization of modified Q-learning with the improved version of particle swarm optimization (IPSO) and arithmetic optimization algorithm (AOA). In the current scenario, classical Q-learning is modified through a reward policy and generates the best solution for PSO. The basic PSO is upgraded through the perception of ascendency in human civilization and generates an optimal location in the succeeding iteration using an arithmetic optimization algorithm. The proposed hybrid algorithm primarily highlights evaluating the optimal deadlock and starvation free subsequent positions of every robot from their current position, optimizing the path distance for every robot. The authentication of the projected hybrid procedure has been confirmed through benchmark function, computer real robot through webbots simulator, and simulation. Further, the efficacy of the projected procedure has been confirmed by equating the result achieved from MQL–IPSO–AOA with Q-learning, AOA, and IPSO, and also equating the result of the projected procedure with state-of- arts.