A mathematical programming approach for walking-worker assembly systems Emre Cevikcan Department of Industrial Engineering, Istanbul Technical University, Istanbul, Turkey Abstract Purpose – It has become increasingly critical to design and maintain flexible and rapid assembly systems due to unpredictable and varying market conditions. The first stage of developing such systems is to restructure the existing assembly system. After designing the manufacturing system, efforts should be made for capacity adjustments to meet the demand in terms of allocating tasks to workers. Walking-worker assembly systems can be regarded as an effective method to achieve flexibility and agility via rabbit chase (RC) approach in which workers follow each other around the assembly cell or line and perform each task in sequence. In this paper, a novel mathematical programming approach is developed with the aim of integrating RC in assembly processes. Therefore, this study is thought to add value to industrial assembly systems in terms of effectively raising engineering control for task allocation activities. Design/methodology/approach – Two consecutive mathematical models are developed, since such a hierarchical approach provides computational convenience for the problem. The initial mathematical programming model determines the number of workers in each RC loop for each segment. In addition, the number of stations and the distribution of station times in the segments is essential. Therefore, the succeeding mathematical programming model generates stations in each segment and provides convenience for the workflow in RC loops. The output of mathematical programming models are the parameters of simulation model for performance assessment. Findings – The effectiveness of the proposed approach was validated by an application in a real-life chair production system. The application resulted in performance improvements for labour requirement (12.5 per cent) and production lead time (9.6 per cent) when compared to a classical assembly system design (CASD) where one stationary worker exists in each station. In addition, it is worth to note that RC leads to a reduced number of workers for a considerable number (39.4 per cent) of test problems. What is more, input as well as output factors have been determined via discriminant analysis and their impacts to the utilization of RC were analyzed for different levels. Practical implications – This study is thought to add value to the industry in terms of effectively providing convenience during production planning and task allocation in assembly lines and cells. Originality/value – To the best knowledge of the author, optimization models for RC considering a real industrial application have not yet been developed. In this context, this paper presents an approach which models RC by the use of mathematical programming in manual assembly processes to address this research gap. The contribution of the paper to the relevant literature is the development of hierarchical mixed integer linear programming models to solve RC problem for the first time. Keywords Mathematical programming, Office chair production, Rabbit chase, Walking-worker assembly systems Paper type Research paper 1. Introduction In recent years, major changes in the world of goods manufacturing have occurred such as worldwide competition, advanced manufacturing-process technologies and new manufacturing system structures. Manufacturing enterprises should design and control their production systems so that they can quickly and economically adapt to unpredictable conditions such as varying volumes and costs (Shewchuck, 2008). Increasing product variety, demand uncertainty and the existence of cost awareness make inventory holding an inappropriate strategy for providing quick response to customers. Therefore, it is essential to shorten production lead time for improving delivery performance by lean production, which considers inventory and non-value added times as waste in production systems (Monden, 1993). Lean manufacturers arrange equipment and workstations in a sequence that supports a smooth flow of materials and components through the process, with minimal transport or delay (EPA, 2007). Manufacturing and/or assembly cells (ACs) are frequently adopted as facility layout alternative for medium volume production/assembly in lean manufacturing systems. In cellular manufacturing, a group of heterogeneous machines and workers are dedicated to producing and/or assembly a family of similar parts (Askin and Huang, 2001). Since workers have an important role in doing tasks for products, the assignment of workers to cells becomes a crucial factor for fully utilization of cellular manufacturing systems (Mahdavi et al., 2012). In addition, mixed model assembly line (MMAL) is utilized for high volume assembly in lean-leveled production systems. Assembly tasks are assigned to an ordered sequence of stations, where each worker repeat just some subset of tasks without movements, such that the precedence relations among the tasks are not violated and the desired performance measures are satisfied. The assembly of complex products with different models are performed with high level of productivity despite of the fact that throughput rigidity, idle times for operators exist in MMALs (Scholl, 1999). Following cell and line design, The current issue and full text archive of this journal is available at www.emeraldinsight.com/0144-5154.htm Assembly Automation 34/1 (2014) 56–68 q Emerald Group Publishing Limited [ISSN 0144-5154] [DOI 10.1108/AA-07-2013-067] 56