Geri Dön

Gri karar verme modelleri ile bulut hizmet sağlayıcısı seçimi

Selection of cloud service providers using grey decison-making models

PDF İndir

  1. Tez No: 1022083
  2. Yazar: GÜLCE TUNALI
  3. Danışmanlar: DOÇ. DR. TUNCAY ÖZCAN
  4. Tez Türü: Yüksek Lisans
  5. Konular: İşletme, Business Administration
  6. Anahtar Kelimeler: Bulut bilişim, DEMATEL, Gri sistem teorisi, Çok kriterli karar verme, Cloud computing, DEMATEL, Gray system theory, Multi criteria decision making
  7. Yıl: 2026
  8. Dil: Türkçe
  9. Üniversite: İstanbul Teknik Üniversitesi
  10. Enstitü: Lisansüstü Eğitim Enstitüsü
  11. Ana Bilim Dalı: İşletme Mühendisliği Ana Bilim Dalı
  12. Bilim Dalı: İşletme Mühendisliği Bilim Dalı
  13. Sayfa Sayısı: Belirtilmemiş.

Özet

Dijital dönüşüm sürecinin hız kazanmasıyla birlikte bulut bilişim teknolojileri kurumların bilgi teknolojileri altyapılarını yönetme biçiminde önemli değişimlere yol açmıştır. Bulut bilişim, veri depolama, işlem gücü ve yazılım hizmetlerinin internet tabanlı platformlar aracılığıyla sunulmasını sağlayarak işletmelere daha esnek, ölçeklenebilir ve maliyet etkin çözümler sunmaktadır. Geleneksel bilgi teknolojileri altyapılarında gerekli olan yüksek donanım yatırımları ve bakım maliyetleri bulut bilişim sayesinde önemli ölçüde azaltılabilmekte, bu durum hem özel sektör hem de kamu kurumları açısından bulut tabanlı çözümlerin kullanımını giderek yaygınlaştırmaktadır. Bulut bilişim teknolojilerinin yaygınlaşmasıyla birlikte kurumlar için en uygun bulut hizmet sağlayıcısının belirlenmesi önemli bir karar problemi haline gelmiştir. Günümüzde küresel hizmet sağlayıcılarının yanı sıra, yerel hizmet sağlayıcıları da farklı hizmet modelleri ve teknolojik altyapılar sunmaktadır. Ancak bu alternatifler arasından en uygun sağlayıcının belirlenmesi yalnızca teknik özelliklere dayalı bir değerlendirme ile gerçekleştirilememektedir. Maliyet, sistem performansı, güvenlik ve gizlilik ile hizmetlerin esnekliği gibi birçok kriterin birlikte değerlendirilmesi gerekmektedir. Bu durum bulut hizmet sağlayıcı seçimi problemini çok kriterli bir karar verme problemi haline getirmektedir. Karar verme sürecinde kullanılan değerlendirmeler çoğu zaman kesin sayısal verilere dayanmayıp uzman görüşleri ve belirsizlik içeren değerlendirmeler içermektedir. Ayrıca değerlendirme kriterleri arasında çeşitli etkileşimler bulunabilmekte ve bu durum karar probleminin analizini daha karmaşık hale getirmektedir. Bu nedenle, söz konusu karar probleminin analizinde, belirsizliği ve kriterler arasındaki ilişkileri dikkate alabilen yöntemlerin kullanılması önem taşımaktadır. Bu çalışmada, bulut hizmet sağlayıcı seçimi problemi gri sistem teorisine dayalı çok kriterli karar verme yöntemleri kullanılarak ele alınmıştır. Çalışmanın ilk aşamasında Gri DEMATEL yöntemi kullanılarak değerlendirme kriterleri arasındaki ilişkiler analiz edilmiş ve kriterlerin önem dereceleri belirlenmiştir. Bu kapsamda maliyet, performans, güvenlik ve gizlilik ile esneklik olmak üzere dört ana kriter dikkate alınmıştır. DEMATEL yöntemi sayesinde kriterler arasındaki neden-sonuç ilişkileri ortaya konulmuş ve karar sürecinde hangi kriterlerin daha etkili olduğu belirlenmiştir. Çalışmanın ikinci aşamasında ise Gri TOPSIS yöntemi kullanılarak alternatif bulut hizmet sağlayıcıları değerlendirilmiş ve sıralanmıştır. Bu aşamada üç global ve iki yerel bulut hizmet sağlayıcı alternatifi ele alınmıştır. Gri TOPSIS yöntemi, alternatiflerin ideal çözüme olan yakınlığı ve negatif ideal çözüme olan uzaklığı esas alınarak en uygun alternatifin belirlenmesine olanak sağlamaktadır. Çalışmada ayrıca elde edilen sonuçların güvenilirliğini değerlendirmek amacıyla duyarlılık analizi gerçekleştirilmiştir. Duyarlılık analizi kapsamında kriter ağırlıklarında meydana gelebilecek değişimlerin alternatif sıralamaları üzerindeki etkileri incelenmiş ve önerilen modelin farklı senaryolar altında ne ölçüde tutarlı sonuçlar ürettiği değerlendirilmiştir. Bu çalışma sonucunda bulut hizmet sağlayıcılarının belirlenen kriterler çerçevesinde sistematik bir şekilde değerlendirilmesine olanak sağlayan bir karar destek yaklaşımı sunulmuştur. Çalışmanın bulguları hem akademik literatüre katkı sağlamakta hem de bulut bilişim altyapısına geçiş sürecinde olan kurumlar için uygulanabilir bir değerlendirme çerçevesi sunmaktadır.

Özet (Çeviri)

The continuous advancement of information and communication technologies has fundamentally transformed the way organizations manage, process, and utilize information resources. In recent years, digital transformation has become a strategic priority for organizations operating in both the public and private sectors, leading to an increasing demand for flexible, scalable, and cost-effective information technology solutions. Within this context, cloud computing has emerged as one of the most significant technological developments by enabling organizations to access computing resources, storage capacity, software applications, and network services over the Internet without relying on extensive physical infrastructure. Rather than investing in expensive hardware, software, and maintenance activities, organizations can obtain the required computing resources on demand and pay only for the services they use. This service-oriented approach provides considerable advantages in terms of operational efficiency, scalability, business continuity, and technological flexibility. Consequently, cloud computing has become an indispensable component of modern information technology strategies and an important driver of organizational digital transformation. The widespread adoption of cloud computing has resulted in the emergence of numerous cloud service providers offering a broad range of infrastructure, platform, and software services. Although the availability of multiple alternatives provides organizations with greater flexibility, it also makes the provider selection process considerably more complex. Global cloud service providers offer extensive service portfolios supported by large-scale infrastructures and advanced technological capabilities. At the same time, local cloud service providers provide services that may better satisfy local legal regulations, language requirements, customer support expectations, and data residency policies. Since every organization possesses different operational objectives, financial limitations, security expectations, and technical requirements, identifying the most appropriate cloud service provider represents a strategic decision that directly influences organizational performance and long-term competitiveness. Cloud service provider selection cannot be evaluated solely from a technical or financial perspective. Instead, organizations must simultaneously consider numerous quantitative and qualitative criteria while comparing available alternatives. Factors such as service costs, system performance, scalability, security, privacy, service availability, integration capabilities, and customization options should all be evaluated together to ensure that the selected provider satisfies organizational expectations. Furthermore, these evaluation criteria frequently influence one another. For example, increasing security requirements may lead to higher implementation and operational costs, while enhanced system performance may require more sophisticated infrastructure investments. Likewise, greater flexibility and customization capabilities may increase implementation complexity. These interdependencies indicate that cloud service provider selection should be regarded as a complex multi-criteria decision-making (MCDM) problem rather than a conventional selection process. Another important challenge associated with cloud service provider selection is the existence of uncertainty throughout the evaluation process. In real-world decision environments, decision makers rarely possess complete, objective, or perfectly measurable information regarding every evaluation criterion. Instead, they frequently rely on expert knowledge, professional experience, subjective judgments, and linguistic evaluations. Consequently, the available information generally contains uncertainty and incompleteness, making it difficult to represent evaluations using precise numerical values. Traditional decision-making methods based exclusively on deterministic data may therefore fail to capture the uncertainty inherent in practical decision problems. This limitation has encouraged researchers to employ uncertainty-based decision-making approaches capable of representing imprecise information more effectively. Among these approaches, Grey System Theory provides a practical framework for modeling uncertainty without requiring extensive probabilistic assumptions or complex membership functions. Unlike fuzzy set theory, which requires the definition of membership functions, grey theory represents uncertain information using interval-based grey numbers defined by lower and upper bounds. This characteristic makes grey theory particularly suitable for decision problems where only limited information or expert judgments are available. By preserving uncertainty throughout the computational process, grey methods enable decision makers to obtain more realistic and reliable evaluation results under incomplete information conditions. Considering these characteristics, this thesis proposes an integrated decision-making framework combining Grey DEMATEL and Grey TOPSIS for the evaluation and selection of cloud service providers. The proposed methodology aims to identify the relationships among evaluation criteria, determine their relative importance, and rank alternative cloud service providers while explicitly considering uncertainty in expert evaluations. By integrating these two methods, the study establishes a comprehensive decision support framework capable of handling both criterion interdependencies and uncertain decision environments. The first stage of the proposed methodology employs the Grey DEMATEL method to analyze the causal relationships among evaluation criteria. DEMATEL is widely recognized as an effective technique for identifying complex interactions within decision systems because it distinguishes between influencing and influenced criteria. Instead of assuming that evaluation criteria are independent, DEMATEL explicitly models their mutual relationships and quantifies the degree to which each criterion affects the others. Incorporating grey numbers into the DEMATEL procedure further enables the analysis to accommodate uncertainty associated with expert judgments. Within the scope of this research, four main evaluation criteria are identified through an extensive literature review and expert opinions: cost, performance, security and privacy, and flexibility. In order to perform a more detailed assessment, each main criterion is represented by three sub-criteria. The cost criterion includes usage cost, scaling cost, and total cost of ownership. The performance criterion consists of latency, availability, and load capacity. Security and privacy are represented through data privacy, certifications and standards, and authorization and access management. Finally, the flexibility criterion includes scalability, integration capability, and customization. This hierarchical evaluation structure provides a comprehensive representation of the factors influencing cloud service provider selection. The Grey DEMATEL analysis determines both the cause-effect relationships among the criteria and their relative importance within the overall decision framework. By calculating the total relation matrix and deriving the prominence and relation values, the method identifies which criteria exert greater influence over the decision process and which criteria are primarily affected by other factors. Consequently, the obtained criterion weights reflect not only the individual importance of each criterion but also their interactions within the evaluation system. This characteristic distinguishes DEMATEL from conventional weighting techniques that assume criterion independence. After determining the criterion weights through Grey DEMATEL, the second stage of the proposed framework applies the Grey TOPSIS method to evaluate and rank the alternative cloud service providers. The TOPSIS methodology assumes that the most desirable alternative should simultaneously possess the minimum distance from the positive ideal solution and the maximum distance from the negative ideal solution. Integrating grey numbers into the TOPSIS procedure enables uncertain expert evaluations to be preserved during the ranking process, thereby increasing the reliability of the obtained results under incomplete information conditions. Five cloud service provider alternatives are evaluated within the scope of the application. These alternatives consist of three global cloud service providers and two local cloud service providers. Expert evaluations are collected for each alternative according to the defined evaluation criteria, after which grey decision matrices are constructed and normalized. The weighted normalized decision matrix is subsequently generated by incorporating the criterion weights obtained from the Grey DEMATEL analysis. Finally, the separation measures from the positive and negative ideal solutions are calculated, and the relative closeness coefficients are determined to establish the final ranking of the cloud service providers. The application results demonstrate the effectiveness of the proposed integrated methodology in evaluating cloud service providers under uncertain decision conditions. According to the Grey DEMATEL analysis, performance is identified as the most influential main criterion within the evaluation framework, indicating that technical performance characteristics play a dominant role in cloud service provider selection. Cost, flexibility, and security and privacy also make substantial contributions to the decision process, emphasizing that organizations should evaluate cloud service providers from multiple perspectives rather than relying on a single criterion. The Grey TOPSIS analysis ranks Global Cloud Service Provider-2 as the most appropriate cloud service provider among the evaluated alternatives. Global Cloud Service Provider-1 is identified as the second-best alternative, followed by Global Cloud Service Provider-3. Local Cloud Service Provider-1 and Local Cloud Service Provider-2 occupy the fourth and fifth positions, respectively. These findings indicate that globally established cloud service providers demonstrate superior overall performance when all evaluation criteria are considered simultaneously. Nevertheless, the results also highlight that local providers may remain competitive under specific organizational requirements, particularly when local regulatory compliance or operational considerations are prioritized. In addition to the primary evaluation, a sensitivity analysis is conducted to investigate the robustness and stability of the proposed decision model. Sensitivity analysis represents an essential component of multi-criteria decision-making studies because small variations in criterion weights may influence the final ranking of alternatives. In this study, different weighting scenarios are generated by modifying the importance levels of the main evaluation criteria while maintaining the overall consistency of the weighting structure. The ranking results obtained under these alternative scenarios are then compared with the original ranking in order to assess the stability of the proposed framework. The sensitivity analysis indicates that the proposed Grey DEMATEL–Grey TOPSIS methodology produces stable and reliable ranking results under different weighting conditions. Although minor changes occur in the relative closeness values when criterion weights are modified, the overall ranking structure remains largely unchanged. This outcome demonstrates that the proposed model is robust against reasonable variations in expert evaluations and criterion importance, thereby increasing confidence in the practical applicability of the obtained results. The contributions of this thesis can be evaluated from both methodological and practical perspectives. From a methodological standpoint, the study demonstrates the successful integration of Grey DEMATEL and Grey TOPSIS within a unified decision-making framework capable of simultaneously addressing criterion interdependencies and uncertainty. The proposed methodology provides an effective approach for solving technology selection problems characterized by incomplete information and subjective expert judgments. Furthermore, the study enriches the existing literature on cloud service provider selection by presenting an integrated grey multi-criteria decision-making model supported by a comprehensive sensitivity analysis. From a practical perspective, the proposed decision framework offers organizations a systematic, transparent, and analytically rigorous tool for evaluating cloud service providers. Decision makers can utilize the methodology to compare alternative providers according to their organizational priorities while considering both technical and managerial evaluation criteria. The proposed framework supports more objective decision-making by reducing the influence of purely subjective judgments and providing a structured evaluation process. In conclusion, this thesis demonstrates that integrating Grey DEMATEL and Grey TOPSIS constitutes an effective and reliable approach for cloud service provider selection under uncertain decision environments. By combining causal relationship analysis, objective criterion weighting, alternative ranking, and sensitivity analysis within a single framework, the proposed methodology provides a comprehensive decision support model that contributes to both academic research and practical decision-making. The findings may serve as a useful reference for organizations planning cloud adoption strategies and for researchers seeking to develop advanced multi-criteria decision-making models for technology evaluation problems.

Benzer Tezler

  1. Tez No

    495245

    Çok kriterli karar verme yöntemleri kullanılarak personel seçim kriterlerinin ağırlıklandırılması ve terfi için personel seçimi

    Weighting personnel selection criteria and personnel selection for promotion by using multi criteria decision making methods

    KEMAL GÖKHAN NALBANT

    Yüksek Lisans

    Türkçe

    Türkçe

    2017

    Endüstri ve Endüstri MühendisliğiYıldız Teknik Üniversitesi

    Endüstri Mühendisliği Ana Bilim Dalı

    Prof. Dr. HÜSEYİN BAŞLIGİL

  2. Tez No

    765867

    Gri sistem teorisi ve bir uygulama

    Grey system theory and an application

    CEREN SÜMER

    Yüksek Lisans

    Türkçe

    Türkçe

    2022

    EkonometriAnkara Hacı Bayram Veli Üniversitesi

    Ekonometri Ana Bilim Dalı

    PROF. DR. MURAT ATAN

  3. Tez No

    938348

    Ar-Ge projelerinde risk ve performans yönetiminde ödül-ceza modelleri ve hibrit çok kriterli karar verme karşılaştırılması

    Comparison of reward-penalty models and hybrid multi-criteria deci̇si̇on maki̇ng i̇n R&D project ri̇sk and performance management

    EZGİ GÜL DİNÇER

    Yüksek Lisans

    Türkçe

    Türkçe

    2025

    Endüstri ve Endüstri MühendisliğiSakarya Üniversitesi

    Mühendislik Yönetimi Ana Bilim Dalı

    DOÇ. DR. SAFİYE SENCER

  4. Tez No

    812012

    Erken uyarı sistemleri ve çok kriterli karar verme yöntemlerine göre finansal performans analizi: Borsa İstanbul üzerine bir uygulama

    Financial performance analysis according to early warning systems and multi-criteria decision making: An application on Borsa Istanbul

    İREM KARAPINAR

    Yüksek Lisans

    Türkçe

    Türkçe

    2023

    İşletmeKocaeli Üniversitesi

    İşletme Ana Bilim Dalı

    DOÇ. DR. HAKAN KAPUCU

  5. Tez No

    935937

    Yapay zeka optimizasyon algoritmaları ile doğrusal olmayan regresyon modellerinde parametre tahmini

    Parameter estimation of nonlinear regression models with artificial intelligence optimization algorithms

    MURAT AÇIKGÖZ

    Yüksek Lisans

    Türkçe

    Türkçe

    2025

    İstatistikAnkara Üniversitesi

    İstatistik Ana Bilim Dalı

    PROF. DR. ÖZLEM TÜRKŞEN

Geri Dön