چكيده به لاتين
Nowadays, many industries and companies need continuous and high-quality production in such a way that, while estimating the demand, they also manage an economical system. Since the continuation of production is dependent on the uninterrupted operation of machines, the optimal management of the supply and inventory of equipment and spare parts guarantees the continuation of the life of production; Therefore, a supply chain that can achieve the highest level of response and the lowest cost, considering the environmental uncertainty, will be desirable for industries. This research has defined the proposed problem in the form of network design and planning by examining the literature and the case study of the Iranian oil industry. In this research, the mathematical models of network design, planning, and forecasting have been developed with the perspective of the integration of decisions, so that with the optimal management of inventory and the optimal allocation of repairs to the repair centers, the costs are reduced and the performance of the chain is improved. Also, according to the variable demand over time, using Poisson and Weibull distributions, demand estimation has been done according to the demand pattern of each spare part. Location decision is considered in the network design model and the planning model concerns the decisions related to, reorder points, inventory level, and order allocation to suppliers and repair centers are determined. Also, considering the value of data, the data-driven consolidation model has been developed. To optimize the model in problems with large dimensions, a heuristic method has been used and the validation of the solution method has also been done. The results show that the integration of decisions reduces the prediction error and, as a result, increases the accuracy and performance of planning. In addition, using the piecewise linearization method in the multi-period model optimizes the costs and the level of availability, while the use of the developed models is a well-structured framework for decision-making regarding the purchase and repair of spare parts. The results of the data-driven model show the reduction in error in the case of using regression through the piecewise estimation method by considering the data variance in addition to the mean.
