Scheduling dynamic cellular manufacturing systems in the presence of cost uncertainty using heuristic method

Nowadays in production rivalry world, designing appropriate layouts and locating machines is an important step in lean production which can increase the performance of a manufacturing system. Cellular Manufacturing Systems covers a wide range of industries. Cost uncertainty which can happen as a result of market changes and inflation rate, is a big concern in scheduling cellular systems as they may impose financial harms to a manufacturing system. Initial investigation in the literature of the CMS studies reveals that the issue uncertain cost in CMS is less developed. It is assumed that a strong production planning can smooth the consequences of uncertain costs in CMS. In this regard, 4 mathematical programming models are developed for forming cells and scheduling the materials on appropriate machines while the system costs are considered uncertain. Since the proposed models (like similar models in the literature) are likely to fall into local optimum points, a Branch and Bound based heuristic, a hybrid Simulated Annealing and Genetic algorithm, a hybrid Tabu search and Simulated Annealing, a hybrid Genetic algorithm and Simulated Annealing, a hybrid Ant Colony Optimization and Simulated Annealing and a hybrid Multi-layer Perceptron and Simulated Annealing algorithms are developed. Then, design of experiments is used to examine the sensitivity of the parameters of each solving algorithm using Taguchi method. Afterward, the proposed solving methods are verified using 17 data sets from the literature and results are analyzed. Results show that during the part-routing process in a normal manufacturing circumstance, nearer set of required machines for producing a product are employed more than other parallel machines.This phenomenon can cause increasing machine-loads in such cells and may lead to machine-load variation in set of closer machines while other machines are allocated less (or even left idle). It is observed that in 67% of studied cases, inflation rate can strengthen cell load variation. To prevent this event a new method is proposed using the statistical process control terms (SPC) which prevent allocating each machine type more than a dynamic upper limit of average of a machine type inside a cell. Results show that in 96.7% of studied cases, the proposed method can significantly prevent machine over allocating in cellular manufacturing systems. While machine broken comes into account, it is found that machine unreliability can cause increasing machine-load variation and strengthen the system imbalance as well. It is also found that using appropriate preventive maintenance program can cause up to 75% reduction in cell-load variation. Similarly using a proper plan for promoting human resources can significantly reduce cell load variation (76% of studied cases). After designing a cellular manufacturing system and during constructing period, it is shown that using an appropriate backward method to maximize the Net Present Value of activities can be used as a tool for reducing the financial harms imposed by uncertain costs