An Improved population-based Simulated Annealing algorithm for solving a Green Vehicle Routing Problem

Within the realm of sustainable development, the Green Vehicle Routing Problem (GVRP) has attracted considerable attention due to its capacity to optimize routes to minimize emissions within third-party logistics. As an NP-hard combinatorial problem, the GVRP demands the utilization of various metaheuristic algorithms. Nevertheless, these algorithms are often crafted to handle continuous decision variables and necessitate adaptation to accommodate binary/integer variables. This investigation addresses the Emissions Vehicle Routing Problem (EVRP), an extension of the GVRP, with the objective of minimizing both distance traveled and carbon dioxide (CO
2
) emissions. To this end, we propose an enhanced population-based simulated annealing (IPSA) algorithm. The IPSA is crafted by combining a novel technique for generating feasible neighboring real-valued solutions, thereby facilitating a profound exploration of the search space. Furthermore, we introduce a novel encoding/decoding methodology that translates a prospective realvalued solution into a viable integer-valued solution for the EVRP dilemma. Through exhaustive experimentation and comparative search across various EVRP instances, our findings evince the efficacy of the IPSA algorithm regarding to well-known metaheuristic algorithms. Our inquiry underscores the capacity of the IPSA algorithm to attain high-caliber EVRP solutions.