الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 191 |  |  |
| | | from | 191 | | |
Abstract
Nowadays, autonomous warehouses become essential for large companies that serve millions of users. One of the most challenging problems inside warehouses is the order picking routing problem, which can be formulated as a traveling salesman problem (TSP). The main objective of this thesis is to solve this problem using a new variant of the Cuckoo Search Algorithm. The work in this thesis is achieved via set of phases. In phase one, three algorithms are proposed to make Levy flight step size parameter adaptive. Levy flight step size is responsible for the size of modification in the solutions. The first algorithm, called Damped Cuckoo Search (DCS), is based on the damped oscillations of the second order systems, while the other two algorithms are based on the chaotic maps. The proposed algorithms are compared to other ten adaptive CSA variants on CEC2017 benchmark functions. The results proved the superior performance of the DCS Algorithm. In phase two, a novel algorithm, called Doubly Exponential Cuckoo Search (DECS), is proposed to make the probability of discovery parameter adaptive. The probability of discovery is responsible for the balance between the exploration and exploitation processes. This algorithm is based on the double Mersenne numbers. The proposed algorithm is compared to other nine adaptive CSA variants on CEC2017 benchmark functions. The results proved the superior performance of the DECS algorithm. In phase three, a new Discrete variant of the Cuckoo search Algorithm, called Discrete Damped Cuckoo Search (DDCS), is proposed to solve the Travelling Salesman Problem. Various modifications are made on the standard CSA. This algorithm is based on using adaptive CSA parameters and the random key encoding. The solution is updated using the 2-opt moves for small changes, and the L\’evy flight random walk for large changes. In case of discovery, the population is divided into sub-groups, then different mutation techniques are applied on each sub-group. In phase four, the proposed DDCS algorithm is customized to fit the warehouse environment. The proposed DDCS is compared to the Genetic Algorithm on set of common TSP benchmark problems and on different path planning problems. The statistical analysis proved that the proposed DDCS algorithm outperformed the Genetic algorithm. Additionally, simple warehouse prototype is implemented to test the functionality of the proposed algorithm. The practical results are as expected by the simulation results.