Demand forecasting is a key factor for apparel retail stores to sustain their business, especially where there are variety of products and intermittent demand. In this study, two of the most popular machine learning methods, random forest (RF) and k-nearest neighbour (KNN), have been used to forecast retail apparel’s intermittent demand. Numerous variables that may have an effect on the sales, have been taken into account one of which is defined as “special day” that might trigger intermittence in the demand. During the application of the forecast, four different datasets were used to provide reliability. 28 different variables were used to increase accuracy of the forecasting and experience of the behaviours of the algorithms. Root mean square error (RMSE) was used to evaluate performance of the methods and as a result of this study, RF showed better performance in all four datasets comparing to KNN.
10 March 2022 in Journal of Advanced Research in Natural and Applied Sciences
Günümüzde; doğal afetlerin artması, teknolojinin hızla gelişmesi, ürün ömrünün kısalması, maliyetlerin artması, politik sebepler, arz ve talepteki belirsizlikler sonucu giderek artan tedarik zinciri riskleri ile karşılaşılmaktadır. İşletmeler rekabet avantajı sağlayabilmek, müşteri beklentilerini karşılayabilmek, sektörde pazar paylarını artırabilmek ve sürdürülebilirliklerini koruyabilmek için tedarik zinciri risklerini sistemli bir şekilde yönetmek durumundadırlar. Çalışmada, iklimlendirme ve ısıtma sektöründeki bir firmanın talep planlama sürecinde oluşan hatalar analiz edilmiştir. Bu çalışmada amaçlanan; Süreç HTEA (Hata Türü ve Etkileri Analizi) yönteminin zayıf yönleri sebebiyle, risk analizine Bulanık SWARA-Bulanık COPRAS tabanlı süreç aşamalı yeni bir yaklaşım getirmektir. Süreç HTEA yönteminin, alanında uzman karar vericilerin bilgi ve tecrübelerini değerlendirmelerde göz önüne alamaması şeklindeki zayıflıklarını iyileştirmek hedeflenmiştir. Bu amaçla, risk analizi için Bulanık SWARA-COPRAS tabanlı süreç aşamalı ÇKKV (Çok Kriterli Karar Verme) modeli önerilmiştir. Kriterlerin önem ağırlıklarını belirleyebilmek için Bulanık SWARA yöntemi, hataları sıralayabilmek için Bulanık COPRAS yöntemi kullanılmıştır. Bu yönteme göre, en önemli hataların; “HT-1: Talep rakamlarının eksik veya yanlış olması sonucu eksik üretimin S&OP (Satış ve Operasyon Planlama) toplantısında farkedilmesi” ve “HT5:Talep rakamları belirlenirken master datanın eksik tanımlanması sonucu eksik üretimin S&OP toplantısında fark edilmesi” olduğu tespit edilmiştir. İlerleyen çalışmalarda, risk analizlerinin etkinlikleri benzer yaklaşımlarla değerlendirilebileceği ve karar destek sistemlerine entegre edilerek mevcut veya gelecek senaryoların analizinde kullanılabileceği öngörülmektedir.
15 April 2021 in Sakarya University Journal of Science
Group technology’s basic logic is grouping and producing products of the same type together. An important reason behind Group Technology becoming such an important topic is that nowadays companies have quite an extensive range and workshop type production has increased. Both fuzzy clustering and rank order clustering methods use for grouping parts and machines based on a part-machine matrix created from the production flow technique in order to increase productivity and reduce cost and workmanship required. In this study, Group Technology techniques such as the rank order clustering and fuzzy clustering methods were applied in order to increase the efficiency of the production line, reduce transportation between machines, and form a machine-parts groups in the wood cutting department of a furniture company producing modular furniture in Istanbul. The TOPSIS method was used to determine which products to take into account. According to results of the study, it is shown that fuzzy clustering method has overperformed rank order clustering method based on the evaluation criteria which are group productivity with 21,36%, group efficiency with 43,21% and grouping measure with 82,33%.
Uncoordinated and instantaneous charging decisions made by electric vehicle (EV) drivers create bottlenecks in existing infrastructure, leading to inefficiencies and prolonged waiting times, and resource losses that challenge sustainable transportation systems. This study proposes a “scenario-based” optimization approach targeting the stochastic behaviors of independent EV drivers, incorporating individual risk-taking profiles and balking mechanisms to promote infrastructure sustainability. The proposed algorithm integrates a discrete-event simulation with a Genetic Algorithm (GA) as a decision support mechanism. The optimization focuses on a vehicle cohort entering the route once the system reaches a steady-state saturation point during peak evening hours. GA parameters are optimized using the Taguchi method to maximize robustness. The results demonstrate that, compared to the baseline scenario where drivers act individually, the proposed decision-making mechanism can achieve up to a 20% reduction in the total travel time of the optimized vehicle group. Overall, the proposed model offers a scalable framework for optimizing individual charging behaviors, thereby contributing to more predictable, resource-efficient, and sustainable management of electric vehicle charging infrastructures.