Browsing by Author "Kose, Simge Yelkenci"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Article
    Citation - WoS: 16
    Citation - Scopus: 20
    Capacity Improvement Using Simulation Optimization Approaches: a Case Study in the Thermotechnology Industry
    (Taylor & Francis Ltd, 2015) Kose, Simge Yelkenci; Demir, Leyla; Tunali, Semra; Eliiyi Türsel, Deniz
    In manufacturing systems, optimal buffer allocation has a considerable impact on capacity improvement. This study presents a simulation optimization procedure to solve the buffer allocation problem in a heat exchanger production plant so as to improve the capacity of the system. For optimization, three metaheuristic-based search algorithms, i.e. a binary-genetic algorithm (B-GA), a binary-simulated annealing algorithm (B-SA) and a binary-tabu search algorithm (B-TS), are proposed. These algorithms are integrated with the simulation model of the production line. The simulation model, which captures the stochastic and dynamic nature of the production line, is used as an evaluation function for the proposed metaheuristics. The experimental study with benchmark problem instances from the literature and the real-life problem show that the proposed B-TS algorithm outperforms B-GA and B-SA in terms of solution quality.
  • Thumbnail Image
    Conference Object
    An Integrated Binary-Tabu Search Approach for the Buffer Allocation Problem: An Industrial Case Study
    (Dime Univ Genoa, 2012) Demir, Leyla; Kose, Simge Yelkenci; Tunali, Semra; Eliiyi Türsel, Deniz
    In this study, an integrated binary-tabu search algorithm is proposed to solve the buffer allocation problem for a real manufacturing system producing heating exchanger in Turkey. The aim is to minimize the total buffer size in the system while improving the system performance. To achieve this objective first the production system is modeled by using simulation. After that the proposed algorithm is employed to find the optimal buffer levels for minimizing the total buffer size in the system. The experimental study shows that proposed algorithm improves the current average daily throughput rate about 31.68%.