الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 122 |  |  |
| | | from | 122 | | |
Abstract
The Vehicle Routing Problem (VRP) is one of the most challenging combinatorial optimization tasks. This problem has been identified more than 40 years ago, and it is to design the optimal set of routes for the vehicle fleet to serve a specific group of customers. The motivation behind the interest in VRP is its practical suitability and the great difficulty it faces. The VRP aims to serve a group of customers with known demands on the routes of the lowest-cost vehicles that start and end in a warehouse.
VRP is a known integer programming problem that falls within the category of NP-Hard problems. Usually, this task is accomplished by using different meta-heuristic methods. VRP naturally arises in 1959 as a central problem in the areas of transportation, distribution, and logistics. In some market sectors, transportation means a high proportion of the added value of goods. Usually, in real-world VRPs, many side constraints appear. Some of the most important restrictions are: Every vehicle has a limited capacitate (Capacitated VRP - CVRP).
In this thesis, the Capacitated Vehicle Routing Problem (CVRP) will be presented. The CVRP is the most popular type of vehicle routing problem. The purpose of this problem is to decrease the total distance traveled by vehicles with respect to restrictions of vehicles’ capacity.
There are many theories and algorithms to solve the CVRP, among the most famous algorithms used are meta-heuristic algorithms like (Swarm Algorithms). An example of Swarm algorithms is Particle Swarm Optimization (PSO). In this thesis, the Chicken Swarm Optimization (CSO) algorithm will be applied to solve the CVRP.
After applying the CSO, the results of the CSO algorithm will be compared with the Networking and Emerging Optimization Research group (NEO) benchmark results, and the results exceeded those on the NEO benchmark. But when the results of the CSO compared with the results of the PSO, the results of the PSO were better than the results of the CSO.
For that, an attempt was made to improve the CSO by combining it with other algorithms as a hybrid algorithm. The first proposed hybrid algorithm is the merging of the CSO algorithm with the Tabu Search algorithm (TS) algorithm. The results from a computational experiment show that the hybrid algorithm can be considered as an efficient approach and overcome the NEO benchmark best-known results by 90%, also from the simulation results, the proposed hybrid algorithm overcomes the CSO results by 70%. But also we found that with the improvement of the results that was extracted by the Hybrid CSO with TS (HYCSOTS) algorithm, it was also not better than the results of the algorithm the PSO. For that, in an attempt to improve the results of the HYCSOTS algorithm, the Genetic algorithm (GA) combined with HYCSOTS algorithm in order to get better results than the results of the PSO algorithm. After comparing the results, the results of the second proposed algorithm Hybrid CSO with GA and TS Algorithms (HYCSOGATS) exceeded the PSO results completely.
Finally, this proves our expectation that (HYCSOGATS) is the best and will surpass all the previously used algorithms to solve CVRP.