Dilay Çelebi

An integrated neural network and data envelopment analysis for supplier evaluation under incomplete information

Supplier evaluation and selection are critical decision making processes that require consideration of a variety of attributes. Several studies have been performed for effective evaluation and selection of suppliers by utilizing several techniques such as linear weighting methods, mathematical programming models, statistical methods and AI based techniques. One of the successful evaluation methods proposed for this purpose is data envelopment analysis (DEA), that utilizes techniques of mathematical programming to evaluate the performance of a set of homogeneous decision making units, when multiple inputs and outputs need to be considered. It is often complicated, costly and sometimes impossible to acquire all necessary information from all potential suppliers to attain a reasonable set of similar input and output values which is an essential for DEA. The purpose of this study is to explore a novel integration of neural networks (NN) and data envelopment analysis for evaluation of suppliers under incomplete information of evaluation criteria.

Optimal policies for a delay time model with postponed replacement

We develop a delay time model (DTM) to determine the optimal maintenance policy under a novel assumption: postponed replacement. Delay time is defined as the time lapse from the occurrence of a defect up until failure. Inspections can be performed to monitor the system state at non-negligible cost. Most works in the literature assume that instantaneous replacement is enforced as soon as a defect is detected at an inspection. In contrast, we relax this assumption and allow replacement to be postponed for an additional time period. The key motivation is to achieve better utilization of the system’s useful life, and reduce replacement costs by providing a sufficient time window to prepare maintenance resources. We model the preventive replacement cost as a non-increasing function of the postponement interval. We then derive the optimal policy under the modified assumption for a system with exponentially distributed defect arrival time, both for a deterministic delay time and for a more general random delay time. For the settings with a deterministic delay time, we also establish an upper bound on the cost savings that can be attained. A numerical case study is presented to benchmark the benefits of our modified assumption against conventional instantaneous replacement discussed in the literature.

Multi Criteria Classification for Spare Parts Inventory

Maintenance operations directly influence the performances of railway vehicles and play a crucial role in railway services to provide uninterrupted and high quality service to passengers. With the exception of preventive activities, the demand of spare parts for maintenance tasks is usually random; hence, the fast and secure management of the spare parts inventory is an important factor for the successful execution of the maintenance process. The purpose of this research is to extend the classical ABC analysis by developing a multi-criteria inventory classification approach for supporting the planning and designing of a maintenance system. Relevant classification criteria and control characteristics of maintenance spare parts are identified and selected and discussed in terms of their effects on maintenance operations, purchasing characteristics, positioning of materials, responsibility of control, and control principles.

Light rail passenger demand forecasting by artificial neural networks

The success of strategic and detailed planning of public transportation highly depends on accurate demand information data. Short-term forecasting is the key to the success of transportation operations planning such as time-tabling and seat allocation. This study adopts neural networks to develop short-term passenger demand forecasting models to be used in operational management of light rail services. A multi-layer perceptron (MLP) model is preferred due to not only its simple architecture but also proven success of solving approximation problems. For eliminating the significant seasonality in time slots, each time slot is handled independent of the others, and an artificial neural network based on daily data is developed for each. Regarding to the 74 different time slots, 74 different neural networks are trained by history data. Three illustrative examples are demonstrated on one of the time slots and performance of the forecast models are evaluated based on mean square errors (MSE) and mean absolute percentage errors (MAPE).

Evolutionary Inventory Control for Multi-Echelon Systems

The purpose of this chapter is to present the use of Genetic Algorithm (GA) for solving multi-echelon inventory problems. The literature of GA dealing with inventory control problems is briefly reviewed with particular focus on multi-echelon systems. A novel GA based solution algorithm is introduced for effective management of a stochastic inventory system across a distribution network under centralized control. To demonstrate the performance of proposed GA structure, several test cases with different operational parameters are studied and experimented. The percentage differences between the total cost obtained by GA and the lower bounds and simulation results are used as performance indicators. Findings of the experiments show that the proposed GA approach can be very useful for obtaining feasible and satisfying solutions for the centralized inventory distribution problem.

A Closed-Loop Sustainable Supply Chain Network Design with System Dynamics for Waste Electrical and Electronic Equipment

Supply chain management covers the management of all activities starting from the supply of the raw material to the delivery of the final product to the end user. In the rapidly evolving and globalizing world, limited resources and increasing competitiveness are pushing both nations and organizations to make a difference in the context of supply chain management. The importance of the concept of the sustainability has become more widely recognized among nations and organizations recently. Recovery options are considered to be an economic gain by many companies. Moreover, pricing is no longer a unique competitive strategy since customers give today value and prefer environmentally friendly products. In other words, recovery options are considered by manufacturers due to customer demand, regulations, and economic return. This study puts forward a sustainable supply chain network design with a system dynamics model to minimize the waste of electric and electrical equipment, which is the one of the most crucial sectors in terms of waste management. This study contributes to filling the gap in the literature concerning the mathematical closed-loop reverse supply chain network design model from a system dynamics perspective rather than by using deterministic and static models proposed in the literature. The proposed model is visualized with the program AnyLogic. It constitutes a general framework with crucial variables of electrical and electric equipment supply chains. The results show the states of these variables by year, which can provide a decision-support system for policy making. The study ends by suggesting future directions and giving some helpful recommendations for other researchers on this topic.