DEVELOPMENT OF AN OPTIMIZED ROUTING SCHEME FOR A CAPACITATED VEHICLE MODEL

By

Author

Presented To

Department of Engineering

ABSTRACT
This dissertation presents the development of an optimized routing scheme for a capacitated vehicle model using Firefly Algorithm (FFA). The conventional model is a formal description involving mathematical equations formulated to simplify a more complex structure of logistic problems. The logistic problems are generalized as the Vehicle Routing Problem (VRP). When the capacity of the vehicle is considered, the resulting formulation is termed the Capacitated Vehicle Routing Problem (CVRP). In a practical scenario, the complexity of CVRP increases when the number of pickup or drop-off points increase making it difficult to solve using exact methods. Thus, researchers have over the years, proposed computational methods for solving CVRP problems. In this research, two scenarios of CVRP were considered. The solid waste management and supply chain for retail distribution. Thirty-six instances and ten instances in the solid waste management and retail supply chain respectively were used in formulating the optimization model. Certain parameters like number of vehicles, number of customers (pickup or drop off locations), capacity of vehicles, quantity of demand, the number of routes and depot position were considered in formulating the model. Some constraints like a vehicle must begin and end at the depot, all demand must be met, a customer is visited just once by a distinct vehicle each time, the demand on each route must not exceed the vehicle capacity were used to guide the model creation. Also some assumptions were made like observing normal road conditions with traffic, customers availability and a reduction in the total route distance inevitably reduces time and cost. Simulation was carried out using MATLAB R2015b and performance was evaluated on the two scenarios using total route distance covered. The simulated results shows a significant improvement occurred on the travelled distance with a slight percentage difference due to the enormous distance covered. The outcome indicates that the developed model had an overall improvement of 6.03% over the Particle Swarm Optimization (PSO) on the solid waste management and a 7.36% over the Best Known Solution (BKS) for the retail supply chain using the total route cost as performance metrics. For the various depot positions considered which are the Random, Optimized, Centered and Eccentric (ROCE), it is observed that the optimized depot position which is determined by this model had a 25% best result for the Instances of the solid waste management, 44.44% over the eccentric position and 77.78% over the centered and random depot placement. This informs that the developed scheme has significantly reduced the total travelled distance in a search space which can be applied to the logistics industry to save cost and time.