الفهرس يوجد فقط 14 صفحة متاحة للعرض العام
 |  | from | 94 |  |  |
| | | from | 94 | | |
المستخلص
Selecting suitable projects from a group of candidate projects considering all constraints and limitations is one of the challenging problems that face any organization. The main purpose of this thesis is to develop an integrated framework to solve the Project Portfolio selection (PPS) problem. In order to achieve this purpose, 3 main models are developed. First, a model is developed to solve the PPS problem considering the set of criteria as the objective functions representing them as a single objective using weighted method; this model considers a deterministic data and does not handle any uncertainty. Then to capture the stochastic factors affecting the project selection decision, a simulation optimization framework is developed. The first step that is implemented in the framework is to enhance the first model integrating simulation with the optimization model. The second step is to formulate the problem as a multi-objective model by representing each criterion as a separate objective. A binary Cuckoo Search (CS) was proposed to solve the single objective representation of the problem with deterministic data. Then a Monte Carlo simulation is combined with the binary Cuckoo Search to solve the problem under uncertainty. Furthermore, fast Non-dominated sorting was integrated with Monte Carlo simulation to solve the multi-objective model. The results of Binary Cuckoo Search are compared to the results from Lingo software (using default parameters), and it outperformed Lingo. Regarding the stochastic single objective model, the modified CS showed superior results to the Genetic Algorithm obtained from1. Moreover, the fast Non-dominated sorting Cuckoo Search showed better results than the multi-Objective Cuckoo Search mentioned