برنامه ریزی تولید سیستم تولید چند سایت با استفاده از مدل سازی ترکیبی: مطالعه موردی از صنعت خودرو

The paper deals with lot sizing and scheduling problem (LSSP) of a multi-site manufacturing system with capacity constraints and uncertain multi-product and multi-period demand. LSSP is solved by an hybrid model resulting from the integration of a mixed-integer linear programming model and a simulation model. The hybrid modelling approach is adopted to test a local as well as a global production strategy in solving the LSSP concerned. The model proposed is applied to a supply chain of a multi-site manufacturing system of braking equipments for the automotive industry. Solutions obtained by the hybrid model under the local or the global production strategy are compared with an actual reference production plan. The approach could help decision making in adopting a cooperative, rather than competitive, production strategy.

Integration and synchronism of information and material flows of manufacturing sites belonging to a supply chain (SC) are characteristics strongly required by modern industrial organizations. Automotive industrial groups are introducing structural changes in their manufacturing systems in order to guarantee optimal trade-off between customer satisfaction and production costs (Haag and Vroom, 1996; Proff, 2000). Outsourcing is a common solution to this problem. According to such a strategic perspective, main contractors involve suppliers from development and engineering phases of new products. Automotive manufacturers became more and more dependent on suppliers. An analysis of vulnerability in the supply chain of a mid-size European car manufacturer (Svensson, 2000) revealed that disturbances appearing at the subcontractors level propagate downstream and upstream in the supply chain. As a consequence, a more in depth level of integration and cooperation among car manufacturers and subcontractors are recommended to avoid or reduce supply chain vulnerability. Several manufacturers integrate successfully their internal process to external suppliers and customers in a single SC (Frolich and Westbrook, 2001). Customer–suppliers relationships need to be continuously updated; strategic knowledge and information tend to be shared among manufacturers in a larger measure than in the past inducing more in depth industrial relationships. Three levels of supply chain integration in formulating scheduling policies are defined in Wei and Krajewsky (2000). A “myopic” policy occurs when the top tier member only considers its internal flexibility costs. An “intermediate” policy considers flexibility costs of members nearest to the top tier of the supply chain. Finally, “total” policy considers flexibility costs extended to all members of a supply chain. A further complexity factor occurs when suppliers are involved into different SCs. According to such a point of view, an integrated approach to supply chain management (SCM) requires a cooperative, rather than competitive, approach among legally independent but economically and strategically dependent subjects of a SC (Christopher and Towill, 2000). Main goal of SCM consists in establishing optimal combination of competition and cooperation considered as a basic feature of inter-firm networks (Pfohl and Buse, 2000). Stochastic variability in product demand as well as in manufacturing and transportation times are being increasingly considered as a major source of uncertainty (Shobrys and White, 2000). In this paper, after reviewing supply chain models available in the literature (Section 2), the authors introduce (Section 3) an industrial case concerning the supply chain of multi-site manufacturing system producing braking equipment components for the automotive industry. The hybrid model proposed, consisting of an analytical and a simulation model, is described in Section 4. The paper focuses on investigating the way demand stochastic variability can differently affect technical and economic performances of the whole production system, in case of both local and global optimization strategy. Results are finally provided in Section 5.

In this paper the authors deal with production planning problem of a multi-site manufacturing system subject to capacity constraints in case of an uncertain, multi-product and multi-period demand. An hybrid model, resulting from the integration of a mixed-integer linear programming (MILP) model and a simulation model, is developed to solve a lot sizing and scheduling problem (LSSP). Manufacturing capacity at each site is affected by machine failures and repairs as well as by sequence dependent setup times. The model is applied to an industrial case study concerning a supply chain producing components for both the market of original braking equipments and for the aftermarket. In both cases, product demand is subject to stochastic variability. The supply chain consists of three manufacturing sites located in Italy. Production planning of sites is mutually dependent since the production plan of a site affects the production plan of the site upstream in the supply chain. Two different production strategies have been investigated. The first one considers each manufacturing site as a stand alone business unit which pursues its own technical–economic goals. The second one searches for a global solution by considering the pool of sites as a single manufacturing system without relaxing local constrains. Results obtained outlines the capabilities of the hybrid model proposed in solving the LSSP formulated in case of both local and global production strategy. An iterative procedure is design to integrate the MILP and the simulation models. The procedure leads, after few iterations, to a good solution among the feasible ones. Solution feasibility is mainly limited by capacity constraints at each site and by consistency constraints between production carried out by a site with demand provided to its upstream site. Solutions obtained by the hybrid model under the local or the global production strategy are compared with an actual annual reference production plan of the supply chain. For privacy concerns all data in the case study are slightly modified. The comparison is based on an overall economic performance measure defined as the sum of setup, holding, and delivery delayed costs. In the case study investigated, the local production strategy allowed a reduction of about 19% of average overall cost respect to the reference actual situation as the former strategy was able to find a solution allowing significant reduction in holding and delayed delivery costs. Simulation also revealed less uncertainty (over 18% reduction in the standard deviation) around the average overall economic performance of the solution identified by the hybrid approach. The comparison between solutions obtained by local and global production strategies stressed out the expected gap between the corresponding economic performances. The global production strategy leaded to better economic performance than the one provided by the local strategy, even if penalties occurred locally. Quantitative results are referable to the case study investigated. On the contrary, the hybrid modelling proposed reveals as a general suitable tool that could support decision making in establishing agreements among supply chain sites to find out production conditions where each firm can reach its own technical–economic goal by adopting a cooperative, rather than competitive, production strategy.