Evaluation of emergency ordering policies at Syncron

University essay from Lunds universitet/Produktionsekonomi

Abstract: The aim of our research has been to identify what prevents Syncron from implementing a single item emergency ordering policy in their inventory management system, GIM. Emergency ordering is defined here as the ability to use two supply sources, a normal one and a faster more expensive emergency source. We further wanted to investigate under what circumstances an emergency ordering policy could be used and what the benefits of such a policy would be in comparison to Syncron´s current solutions. To answer these questions, we first analyzed the company´s single item and multi-item inventory control mechanisms. Second, we searched the academic literature for emergency ordering polices that would be compatible with the existing system, and that would have the potential to reduce the costs while retaining service levels. Third, the chosen models were tested against Syncron´s current single and dual supplier ordering policies through a simulation study. The primary obstacles preventing Syncron from implementing an emergency ordering policy is their multi-item optimization algorithm. The target of this algorithm is to obtain a certain overall service level while minimizing the total stock value. The problems of this, in the context of introducing an emergency ordering policy from academic literature, are primary twofold. First, making a total cost optimization, where the extra cost for the emergency replenishment option is weighted against the reduction in inventory, is not supported. Second, most emergency ordering policies in literature utilize a backorder cost instead of a service level requirement and the translation between these concepts is often unsatisfactory. Given these limitations there were few feasible models to evaluate. A compromise was made to select two cost optimizing policies that could cut costs significantly if Syncron would make larger changes, and two heuristics which were tailor-made to fit the the current situation, but which lacked the cost optimizing feature. The chosen policies were • The model in Song and Zipkin (2009) for items with Poisson demand, • an adaption of Roslings (2002) Lost Sales model for fast and erratic items that is referred to in Axsäter (2006), • a short horizon emergency heuristic with a Cost/Service level ratio limit for items with fast and erratic demand and, • a long horizon emergency heuristic with a Cost/Service level ratio limit for items with fast and erratic demand. The most important result of our simulation study is that a simple single supplier (R, Q) policy outperformed all the non-optimizing policies, including Syncron´s rush ordering heuristic, under the real conditions tested. Our simulation study further showed that both the model by Song and Zipkin (2009) and the lost sales model performed better than both Syncron´s current solutions, given that all requirement for these policies were met. Our tailor-made solutions performed only slightly better than Syncron´s rush ordering heuristics under the real conditions. Our sensitivity analysis showed that, under conditions that are attractive for emergency ordering in general, the short horizon emergency heuristic performed better than both Syncron´s policies. Our recommendation to Syncron is to use more caution before enabling the existing rush ordering heuristic for a customer. In most cases this heuristic will do more harm than good. We further encourage Syncron to investigate whether the tailor-made polices can have more apparent benefits in a multi-item setting, where a Cost/Service level ratio limit can be set for a group of items. This would enable the filtering out of items not suited for emergency ordering.

  AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)