الفهرس 
Chapter 1 : IntroductionOnly 14 pages are availabe for public view
 |  | from | 150 |  |  |
| | | from | 150 | | |
Abstract
Vehicle Routing Problems (VRPs) are Nondeterministic Polynomial-time, hard optimization problems that play a vital role in transportation, distribution and logistics, It represent an important competitive factor in transportation business where making a good planning for the transportation fleet is imperative and contributes in cutting costs and make use of transportation material as efficiently as possible.VRPs are representing a generalized form of the Traveling Salesman Problem (TSP). Its objective is to plan a tour for a fleet of vehicles under several constraints, that tour represents the best routes between scattered locations or customer nodes with minimum travelling cost. Recently, many exact methods have been used to solve the VRPs such as exact algorithms based on linear programming techniques and guided local search. Heuristic techniques have received wide interest in researchers’ effort to solve large scale VRP. Many researchers are still trying to find the best heuristic, meta-heuristic or hybrid heuristic techniques which give a very good approximate solution in a properly running time.The Cluster first – Route second is the base of the methodology that used to solve the vehicle routing problem which consists of two main phases. In the first phase, the objective is to group the closest geographical customer locations together into clusters based on their locations, vehicle capacity constraint and the total customers’ demands. In the second phase, the objective is to generate route with the minimum travelling cost for each cluster within less computational time.The SWeep Algorithm & Nearest Neighbor Algorithm (SW&NNA) is used as a hybrid algorithm. The sweep algorithm is used for clustering phase and nearest neighbor algorithm is used for the generate route phase. The hybrid algorithm is applied in two case studies and tested by Augerat’s and Cordeau’s benchmark datasets and comparing with the best known algorithms. The experimental results show that the route generation phase has needed more enhancement, which is done by using an evolutionary algorithm such the particle swarm optimization. Sweep Algorithm & Particle Swarm optimization (SW&PSO) is the hybrid algorithm that applied in two case studies and the experimental results are compared to the previous hybrid algorithm (SW&NNA).The algorithm evaluation results show that using the PSO instead of NNA in route generation phase provide more improvement in the performance of the route generation phase, where all total travelling distances for all instances in Augerat and Cordeau benchmark datasets were improved. In Augerat’s set A, the traveling distance reduced with a percentage up to 21% and the computational time is reduced with a percentage from 9% up to 14%. In Augerat’s set B, the traveling distance reduced with a percentage up to 14% and the computational time is reduced with a percentage from 5% up to 15%. In Augerat’s set P, the traveling distance reduced with a percentage up to 22% and the computational time is reduced with a percentage from 6% up to 10%. And in Cordeau’s set p and pr, the traveling distance reduced with a percentage up to 18% and the computational time is reduced with a percentage up to 10%. The PSO is really added more enhancement and presented better results than the previous one.