Loading...

Development of a Novel Framework for Manufacturing Service Composition and Matching in Public Cloud Platform Using a Game Theoric Approach

Delaram, Jalal | 2021

577 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 54448 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Houshmand, Mahmoud; Fatahi Valilai, Omid; Ashtiani Mofrad Tehrani, Farid
  7. Abstract:
  8. In recent years, cloud manufacturing as a successful business model has attracted a lot of attentions. Many papers and researches have discussed various aspects of cloud manufacturing. Despite of enourmous works in this area we can categorize them using the view to the manufacturing platform. According to the authors’ researches, almost all of the the researchers looked to the system in a centralized manner with private structure. A structure in which a central unit decides for all and seeks to optimize its intended objective function. In private platforms, a central unit which is usually the owner of the company, makes decision for its all subsideries. But, public platform has a different mechanism. Each person, firm is in charge of managing his production in public platforms. So, in this paper we introduced the philosophy of the public cloud manufacturing and distinguish it from private mode. We also discuss the mechanism of these platforms from the definition and evaluation of services to the problem-solving algorithm and the services matching and composition, and we use game theory models to model and solve problems. The public platform’s game model consists of two categories of players, providers and consumers. Each customer has criteria for collaborating with manufacturers and prioritize them. On the other side, providers have criteria for consumers’ orders and prioritize them accordingly. These priorities will later be the basis for selecting partner pairs of customers and manufacturers. Various algorithms have been developed to create pairs between two sets of players, including Deferred Acceptance, Immediate Acceptance, and Top Trading Cycle algorithms. The solution of each algorithms has properties that affect its performance and efficiency. Other influential factors on the solution of the problem are the proposer impact, resources availability impact, and the prioritization method. In the case study section, we developed a platform for 3D manufacturing of general components. Based on the findings of this dissertation, the best algorithm for selecting pairs should be determined by considering the factors governing the platform. Therefore, the best algorithm for pairs in public cloud platforms where the services of the providers are more than or equal with the customer orders is the Deferred Acceptance algorithm with customers as the proposer, and the Deferred Acceptance algorithm with provider as the proposer, otherwise. The best algorithm for pairs in private cloud platforms where the services of the providers are more than or equal with the customer orders is the Kuhn-Munkres algorithm with customers as the target, and the Kuhn-Munkres algorithm with providers as the target, otherwise. Also in this dissertation, using the novel concept of willingness value, the deferred acceptance algorithm was improved in terms of time performance. According to the experiments of this dissertation, using the concept of the effect of wanting the time performance of the algorithm, it improves by an average of 16.91%. Finally, the leading directions for future research and development of this dissertation are stated
  9. Keywords:
  10. Game Theory ; Cloud Manufacturing ; Nash Equilibrium Point ; Stable Matching ; Public Cloud Platforms

 Digital Object List

 Bookmark

No TOC