Characterizing continuous (s, S) policy with supplier selection using Simulation Optimization

Abstract

A real-world inventory control system, due to its nonlinear, stochastic, time-dependent nature, and the presence of complex interactions between supply chain members, can become quite challenging to optimize and requires a complex model. At this point, the Simulation Optimization (SO) model gains a better understanding of the complex and messy phenomenon of the inventory control of supply chain members. By creating SO models for Distribution Center (DC)s and Suppliers, we wish to present flexible and comprehensible research on the important decision of whether to minimize the differences between total overordering cost and total underordering cost (Model 1) or to minimize the total supply chain cost (Model 2). We also try to point out several important issues: the optimal value of the initial inventory, the reorder point, and the order-up-to level in continuous (s, S) policy for each DC and each Supplier; whether SO models can successfully integrate the supplier selection and continuous (s, S) policy for the supply chain environment; how to apply statistical analysis skills to compare these SO models with a greater level of detail. According to the cost analysis, the total supply chain cost of Model 1 is improved approximately 22% with Model 2. Also, Model 2 is the best one according to quantity-based analysis, order-based analysis, probability-based analysis, and lead-time-based analysis. Model 2 can be successfully applied for the actual situation of the supply chain inventory system and companies can obtain a remarkable amount of saving while increasing their competitive edge.

Simulation Optimization Approach to Periodic Review Inventory Control System With Backorders

“…For Suppliers, the average service level is at least 99%. For the utilized average service level formula, one can refer to Göçken, Dosdoğru, Boru, and Geyik (2017). For supply chain members, minimum initial inventory level is 836 units while maximum initial inventory level is 1821 units.…”

A discrete simulation-based optimization approach for multi-period redeployment in emergency medical services

“…Several factors affect efficiency of the search for optimal solutions in OptQuest, such as the number of controls and their bounds, the number of replications and simulations, the complexity of the objective, and the initial values of the controls. 48 A good choice of the initial values can shorten the time it takes to find an optimal solution. Thus, similar to the iterative method, we used the current location of rescue teams (i.e., VSG or HM) and the same nine random location solutions used in section 4.2 to initialize controls (see Appendix 1).…”

“…46 The processing times (i.e., dispatching time, preparation time, on-scene time, DTR time, and drop-off time) were fitted with empirical distributions because goodness-of-fit tests for theoretical distributions provided low p-values. 47 Finally, particular attention was given to model travel times between bases, scenes, DTR services, and hospitals, as this is a critical component to assess redeployment plans adequately. We first estimated the average travel time along roads of the Val de Marne department at various periods, using the global positioning system (GPS) traces of the SAMU-94 ambulances.…”

Simulation optimization approach for dynamic and stochastic closed loop supply chain network

“…Insert Table 3 and Table 4 The objective functions are defined using equation (8) or equation (9). Definition of the average holding cost, the lost sales cost, the order cost per use, the order processing cost, and the processing cost are summarized in Göçken et al [44]. Note that order processing cost for reverse network includes only the order processing cost rate.…”