A solution to dynamic green vehicle routing problems with time windows using spiking neural P systems with modified rules and learning

Dynamic routing problems involve electronic decision making, which compromises reactivity with the quality of decision making. The time for seeking better decisions comes at the price of a lower reactivity to changes in inputs. This factor is especially important in situations where consumers are calling for a service and a good decision needs to be made as quickly as possible. There are two approaches in solving dynamic vehicle routing (DVRP): to run a state solver every time new requests come in, to construct an initial solution and then to update it each time new information comes. The latter is more commonly used and more flexible. Also, the environment is badly affected by factors like CO₂ emissions; noise; etc. A variant of VRP called green VRP (GVRP) has been formulated in this context. The solution strategies for GVRP are developed to help organizations with alternative fuel-powered vehicles to resolve challenges that arise in conjunction with limited refueling facilities as a result of restricted vehicle driving range. Here, the authors propose a spiking neural P system (SN P)-based model with modified rules and learning in association with firefly optimization (FA) to solve the combined version of GVRP and D VRP with time windows, called DGVRPTW. The SN P system proposed here is a multilayer neural system with embedded potentials and learning facilities which uses the rectified linear unit (reLu) as activation functions. The proposed SN P system is used for geo-location clustering, and the firefly algorithm (FA) is used for route optimization. The proposed SN P system can do predictions accurately when a new customer enters the scenario. The scheme has been tested on medium as well as large-scale instances and analyzed different performance measures such as nature of convergence, utilization rate, solution improvement percentage and dynamic measures. Having applications in image classifications, optimization problems, etc., the proposed system is worthy of future study.