Companies who produce items such as biochemicals, ready concrete, milk, eggs and even DVDs always arrive at an important question: What is the quickest, most effective way to distribute these short shelf life products?
Neil Geismar, an associate professor in the department of Information and Operations Management at Texas A&M University, along with fellow researchers Milind Dawande and Chelliah Sriskandarajah from the University of Texas at Dallas, explored the solutions to these companies’ issues with distribution of perishable products.
The zero-inventory production and distribution problem (ZIPDP) is a common problem encountered in situations in which a product cannot be inventoried because of its short shelf life. Short shelf life is determined by either physical characteristics (such as perishable food items or chemicals) or by the limited duration of market interest (such as magazines, DVDs and electronic games).
Geismar’s research was initiated at the request of the Vice President of Operations at Blockbuster, Inc.’s, distribution center in McKinney, Texas. “He asked us for help to make his system run more efficiently,” Geismar explains. The facility supports 5,600 retail stores via 40 regional hubs, or “pool points.”
“Quick delivery of DVDs to the retail outlets is important because DVDs have an extremely short product life cycle, which can be attributed to the ephemeral nature of most entertainment products, to the release of new titles on DVD every week, and to the release date requirements imposed by the movie studios,” Geismar says.
ZIPDP’s challenge is to coordinate the production and transportation operations so that the total cost of operations is minimized while the product lifetime and the delivery capacity constraints are satisfied. Geismar puts it this way: “The product’s limited lifetime implies that no inventory can be held between production and delivery; hence, the two functions are tightly coupled.”
The researchers examined a popular method of distribution, the pool-point delivery model, which is also known as the “hub-and-spoke” delivery model. In pool-point distribution, the product or service is delivered to dispersed clusters of demand points by first delivering large quantities to pool points (hubs), which are centrally located in their respective clusters. The large quantities of products are then dispersed into small carriers for point-to-point connections (spokes).
This system is used by a wide variety of industries: airlines (large jets fly between hubs, small ones from hubs to other cities), freight distribution, light petroleum products, candy distribution, automobile distribution, and newspaper delivery. Each of these industries requires two steps of distributionâ€”first to the “hubs,” then to the “spokes.”
Combining pool-point distribution and zero-inventory products is no simple feat. Geismar’s research discusses a few real-world examples of this type of distribution, including the production and distribution of ready concrete for the construction of venues for the 2004 Athens Summer Olympics. This example served as a case study for the researchers to investigate when considering the production and distribution of Blockbuster DVDs.
Geismar says the research primarily focused on two objectives relevant in the practice: Minimizing the sum of production and delivery time, and minimizing the total cost for producing and delivering products. The research also examined minimizing mean flow time, minimizing maximum lateness, and minimizing the number of late deliveries in the pool point distribution system.
The research details how these systems operate and provide “efficient algorithms to find optimal schedules for various objectives,” the research states.
The researchers’ paper was the first rigorous study of pool-point distribution in a zero-inventory system. Geismar notes that managers’ supply chain decisions are based on production times, delivery times, the cost to hire each truck, and the cost to increase the production rate, so the results of the study proved effective and beneficial to companies facing distribution issues similar to Blockbuster.
Geismar’s research sums it up this way: “By analyzing overall system cost, we provided managerial insights into how different costs for trucks and for production rates affect the optimum decisions on how many trucks to hire and on which production rate to use for various objectives.”
Categories: Research Notes