الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 179 |  |  |
| | | from | 179 | | |
Abstract
Software reuse involves developing software systems by utilizing existing components instead of starting from scratch. This practice improves the quality of the final product, reduces failure rates, and enhances productivity. Incorporating reusable elements like specifications, designs, and test cases leads to more reliable and easily constructed systems while reducing risks in software development. Component-Based Software (CBS) systems and Services-Oriented Architectures (SOA) are examples of approaches based on software reuse, with CBS utilizing integrated components and SOA combining distributed services. Service-oriented computing (SOC) has gained attention due to its loosely coupled services distributed across multiple locations. The development of large and complex systems relies on the web services composition paradigm, which has garnered significant attention in the field of computing. However, the inherent challenges associated with integration, dependency determination, system complexity, user requests, cycles, redundancy, and scalability, as well as the opaque nature of services, necessitate the creation of a formal model for web service compositions. In the SOC paradigm, the crucial role of fault detection and prediction in minimizing testing costs, improving testing quality, and enhancing the reliability of service compositions are emphasized. Unfortunately, service-oriented systems have received less focus, primarily concentrating on individual web services rather than the compositions themselves. Thus, more emphasis is needed on estimating the number and severity of faults and predicting their occurrence. Consequently, the development of fault recovery techniques has become essential for ensuring reliability and resilience in web service composition.
The thesis extensively addresses challenges in testing SOC systems, emphasizing software reuse paradigms. The proposed software reuse process, involving formal models, fault perspectives, and recovery strategies, aims to optimize fault recovery and enhance SOC system accuracy and resilience. The DSW model reduces modeling time by 75-85%, the MISD model improves SDG construction efficiency by 76%, and accuracy by 86%. The SMFPL-DL model outperforms CNN and RNN models by 5-20%. The FR-SDGO model generates recovery compositions 20-32% more efficiently, and the F2RS-WSC model recommends strategies with less than 3% of parsing time and accuracy between 70-88%, collectively representing significant advancements in SOC systems.