Tedarik zincirlerindeki kırılma risklerinin yapay zekâ yöntemleriyle tahminlemesine yönelik Kızılay İşletmelerinde bir uygulama
Küçük Resim Yok
Tarih
2024
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İnönü Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Son yıllarda küresel ölçekte yaşanmakta olan savaşlar, pandemiler ve doğal afetler gibi durumlar tedarik zincirleri üzerinde önemli bir risk unsuru oluşturmaktadır. İşletmeler etkin bir tedarik zinciri yönetimiyle tedarik zinciri kırılma risklerinin meydana gelme olasılıklarını hesaplayıp, bu risklerin olumsuz etkilerini azaltmayı amaçlamaktadır. Bu kapsamda yapılan çalışmada, işletmelerin tedarik zinciri kırılmaları noktasında, önceden öngörü sahibi olabilmeleri için bir model geliştirilmeye çalışılmıştır. Bu çerçevede geliştirilen modelin uygulaması, tedarik zinciri kırılmalarına yönelik ortak risklerin yönetilmesi için farklı sektörlerde hizmet veren Kızılay İşletmelerinde yapılmıştır. Uygulama çalışmasında farklı sektör ve lokasyonlarda yer alan bu işletmelerin tedarik zinciri üst düzey yöneticileriyle görüşmeler gerçekleştirilmiştir. Yapılan görüşmelerde tedarik zinciri üst düzey yöneticilerine ilgili literatür taraması sonucu elde edilen sektör bazlı riskler sunulmuştur. Görüşmelerde tedarik zinciri üst düzey yöneticilerinin belirlediği riskler, ilgili literatürün taranması ve işletme içi dokümanlara içerik analizi yapılması sonucu elde edilen riskler değerlendirilip işletmelerin yapısına en uygun olanları belirlenmiştir. Üç farklı kaynaktan elde edilen risklerin değerlendirilmesi sonucunda 110 adet ortak risk belirlenmiştir. Riskler, meydana gelme olasılığı ve meydana geldiğinde oluşturabilecekleri etki olmak üzere iki farklı boyutta değerlendirilmiştir. Risklerin olasılık ve etki değerlendirmeleri tedarik zinciri üst düzey yöneticileriyle yapılan odak grup görüşmeleriyle belirlenmiştir. Elde edilen sonuçlara yönelik geliştirilebilecek stratejiler, ilgili literatür taraması ve işletme tedarik zinciri üst düzey yönetici görüşü baz alınarak ortaya konulmuştur. Hangi koşullarda hangi stratejinin uygulanacağına yönelik olarak yapay zekâ yöntemleri kullanılmıştır. Bu yapay zekâ yöntemleri çalışmanın yapısına en uygun olduğu düşünülen bulanık mantık, anfis ve regresyon yöntemleridir. Bulanık mantık yapay zekâ yöntemi, anfis ve regresyon yöntemlerinin matlab programlama dilinde ilgili toolbox modülleri kullanılarak ilgili model çalıştırılmış, işletmeler özelinde ortak risklerin tedarik zincirini kırma olasılığı ve bu olasılığa karşı uygulanması gereken stratejiler (ertele, kontrol, takip, transfer, çeviklik) belirlenmiştir. Bulanık mantık yöntemiyle elde edilen çıktılar, anfis ve regresyon yöntemlerinde gerçekleşen değer olarak kullanılıp anfis ve regresyon yönteminde risklerin tedarik zincirinde kırılma olasılıkları tahmin edilmiştir. Anfis ve regresyon yöntemlerinin uygunluk değerleri çeşitli istatistiki performans ölçütleri kullanılarak karşılaştırılmıştır. Çalışmada kullanılan bulanık mantık yapay zekâ yöntemi, anfis ve regresyon yöntemlerinin matlab programlama dilinde app designer toolbox kullanılarak, kullanıcı arayüz tasarımı oluşturulup dinamik bir sistem modeli geliştirilmiştir. Geliştirilen bu modelin farklı sektörlerde yer alan işletmelerin tedarik zincirindeki kırılma ortak risklerine karşı bir erken uyarı sistemi olarak kullanılması planlanmaktadır. Anahtar Kelimeler: Tedarik Zinciri Risk Yönetimi, Yapay Zekâ, Bulanık Mantık, Anfis, Regresyon
Situations such as wars, pandemics and natural disasters that have been experienced on a global scale in recent years pose a significant risk factor on supply chains. With effective supply chain management, businesses aim to calculate the likelihood of supply chain breakage risks and reduce the negative effects of these risks. In this context, the study attempted to develop a model so that businesses can have foresight at the point of supply chain breaks. The application of the model developed within this framework was made in Red Crescent Enterprises serving in different sectors to manage common risks of supply chain breaks. In the application study, interviews were held with the supply chain senior managers of these companies in different sectors and locations. During the interviews, sector-based risks obtained as a result of the relevant literature review were presented to supply chain senior managers. During the interviews, the risks identified by the supply chain senior managers, the risks obtained as a result of scanning the relevant literature and content analysis of internal business documents were evaluated and the ones most suitable for the structure of the businesses were determined. As a result of evaluating the risks obtained from three different sources, 110 common risks were identified. Risks are evaluated in two different dimensions: the probability of occurrence and the impact they may have when it occurs. The probability and impact assessments of the risks were determined through focus group discussions with supply chain senior managers. Strategies that can be developed for the results obtained have been put forward based on the relevant literature review and the opinion of business supply chain senior managers. Artificial intelligence methods have been used to determine which strategy will be applied under which conditions. These artificial intelligence methods are fuzzy logic, amplifier and regression methods, which are thought to be most suitable for the structure of the study. The relevant model was run using the relevant toolbox modules in the Matlab programming language of the fuzzy logic artificial intelligence method, amplifier and regression methods, and the possibility of common risks breaking the supply chain for businesses and the strategies that should be applied against this possibility (postpone, control, follow-up, transfer, agility) were determined. The outputs obtained by the fuzzy logic method were used as the actual values in the amplifier and regression methods, and the probabilities of the risks breaking in the supply chain were estimated in the amplifier and regression method. The fitness values of the amplifier and regression methods were compared using various statistical performance measures. The fuzzy logic artificial intelligence method, amplifier and regression methods used in the study were used in the Matlab programming language and the app designer toolbox was used to create a user interface design and a dynamic system model was developed. This developed model is planned to be used as an early warning system against common risks of breakage in the supply chain of businesses in different sectors. Keywords: Supply Chain Risk Management, Artificial Intelligence, Fuzzy Logic, Anfis, Regression
Situations such as wars, pandemics and natural disasters that have been experienced on a global scale in recent years pose a significant risk factor on supply chains. With effective supply chain management, businesses aim to calculate the likelihood of supply chain breakage risks and reduce the negative effects of these risks. In this context, the study attempted to develop a model so that businesses can have foresight at the point of supply chain breaks. The application of the model developed within this framework was made in Red Crescent Enterprises serving in different sectors to manage common risks of supply chain breaks. In the application study, interviews were held with the supply chain senior managers of these companies in different sectors and locations. During the interviews, sector-based risks obtained as a result of the relevant literature review were presented to supply chain senior managers. During the interviews, the risks identified by the supply chain senior managers, the risks obtained as a result of scanning the relevant literature and content analysis of internal business documents were evaluated and the ones most suitable for the structure of the businesses were determined. As a result of evaluating the risks obtained from three different sources, 110 common risks were identified. Risks are evaluated in two different dimensions: the probability of occurrence and the impact they may have when it occurs. The probability and impact assessments of the risks were determined through focus group discussions with supply chain senior managers. Strategies that can be developed for the results obtained have been put forward based on the relevant literature review and the opinion of business supply chain senior managers. Artificial intelligence methods have been used to determine which strategy will be applied under which conditions. These artificial intelligence methods are fuzzy logic, amplifier and regression methods, which are thought to be most suitable for the structure of the study. The relevant model was run using the relevant toolbox modules in the Matlab programming language of the fuzzy logic artificial intelligence method, amplifier and regression methods, and the possibility of common risks breaking the supply chain for businesses and the strategies that should be applied against this possibility (postpone, control, follow-up, transfer, agility) were determined. The outputs obtained by the fuzzy logic method were used as the actual values in the amplifier and regression methods, and the probabilities of the risks breaking in the supply chain were estimated in the amplifier and regression method. The fitness values of the amplifier and regression methods were compared using various statistical performance measures. The fuzzy logic artificial intelligence method, amplifier and regression methods used in the study were used in the Matlab programming language and the app designer toolbox was used to create a user interface design and a dynamic system model was developed. This developed model is planned to be used as an early warning system against common risks of breakage in the supply chain of businesses in different sectors. Keywords: Supply Chain Risk Management, Artificial Intelligence, Fuzzy Logic, Anfis, Regression
Açıklama
Anahtar Kelimeler
İşletme, Business Administration
