Q-Learning Based Genetic Algorithm for Solving Pollution Route Planning Problem

As China experiences rapid industrial and urban growth, vehicular emissions from road transport have increasingly contributed to CO2 levels. The Pollution Routing Problem (PRP) is a strategic approach designed to find the most cost-effective and eco-friendly transportation routes. This is achieved by balancing operational expenses with environmental impact, considering variables such as vehicle speed fluctuations, fuel usage, emissions output, and driver compensation, to create a solution that is both economically viable and environmentally responsible. This study introduces a novel approach to tackle the Pollution Routing Problem (PRP) by employing a Q-learning-enhanced genetic algorithm (Q-GA).This approach enhances the efficiency of vehicle routing by leveraging the strengths of both reinforcement learning and genetic algorithms in a synergistic manner, aiming to minimise operational and environmental costs. The core lies in the use of Q-learning reinforcement learning mechanism to guide the parameter selection of the genetic algorithm to efficiently explore and utilise the search space in order to find the optimal solution. Experimental results demonstrate that the Q-GA performs well in solving the PRP problem, converges faster than traditional methods, and provides an effective win-win solution for both environmental and economic benefits.