Re-defining the manufacturing strategy: "the signs of the highway"

This project takes place in the plant of Procter and Gamble (P&G) located in Mechelen. Specifically, the project deals with the detergent tablets packing lines that belong to the mentioned plant. In the last years, this plant has made great efforts to increase its flexibility in order to serve the demands of the market. The plant has succeeded in reducing its reaction time, the average cycle times and the inventory levels. Nevertheless, the company has also experienced a rise in operational costs. Two important causes of these high operational costs are the elevated number of change overs and the generation of scrap. In order to reduce the operational costs, the company has initiated the "LPD superhighway" project that strives for higher commonality in the intermediate products (also known as blends). This project has tickled the mind of the planning department who believes that further improvements are possible if the manufacturing strategy is re-defined. A manufacturing strategy involves a lot elements and decisions. We will not consider all of them in this project. This project is limited to three elements of the manufacturing strategy: Line assignment, sequencing and lot sizing. The analysis of the current operations revealed that there is an imbalance between the inventory costs and the change over costs. This imbalance results in high overall costs. The company keeps low inventory levels and makes pressure over the production process to cope with the changing market requirements. Moreover, there is lack of policies for the allocation of resources and a misuse of the models that guarantee the operational efficiency such is the case of the EPQ (Economic Production Quantity). The study consisted in modeling and optimization of the current operations. The model is a combination of different and simple models such as inventory models, queuing models and an algorithm to compute change over losses. The model has a mean value approach in a monthly basis that evaluates decision regarding line assignment, sequencing and lot sizing. The optimization procedure is based in this model. The optimization procedure consists of an iterative process between three optimization methods: A greedy algorithm for the line assignment, the nearest neighbor for the sequencing and a truncated EPQ model for the lot sizes. The complexity of finding an optimal solution is very high. Therefore, a scenario analysis was used to test different strategies such as brand dedicated lines, size dedicated lines or a flexible allocation. The results of the scenario analysis together with the optimization procedure indicate that the most cost effective strategy is to have brand dedicated lines. The savings of this strategy can reach 300.000 Euros per year in comparison to the current situation. A sensitivity analysis of demand and capacity was also carried out. This analysis shows that the savings can be larger when the number of packing lines is reduced. It is possible to run only 9 packing lines (instead of 11) for the fiscal year 2007-2008. However, this is conditioned by improvements in the production process. Moreover, the results of the model also show that reducing the number of lines while keeping low inventory levels can lead to looses of responsiveness and flexibility. The model was implemented in a excel file with a user friendly interface. The user of this file can vary the parameters and the inputs of the model in order to test new scenarios. The scenarios can be evaluated according to costs and performance measurements of responsiveness, flexibility and use of capacity. The reports of the excel file are used to determine the strategy and policies for the period between January 2008 and June 2008. The capacity analysis showed that it is possible to run 9 lines without diminishing the performance of the plant. Brand dedicated lines, fixed minimized sequences and the EPQ model can lead to savings of around 280.000 Euros for this period.