It takes more than math to create the optimal design for a distribution network in emerging economies. External factors such as regional product demand, commercial real estate prices and transportation costs can change markedly over the three- to five-year planning horizon that is common for such projects. To arrive at the most efficient network design, mathematical optimization models need to be supplemented by analyses of these variable external factors.
Tausif Bashir, a graduate of the MIT-Malaysia Master of Science in Supply Chain Management Class of 2016 at the Malaysia Institute for Supply Chain Innovation (MISI), developed a methodology for taking account of externalities when designing a distribution network. The research work was carried out for his master’s thesis, and was supervised by Dr. Shardul Phadnis, Director of Research, MISI. Wadala Chemicals (a pseudonym), a chemical manufacturer, sponsored the thesis.
Distribution network designs specify the locations of warehouses and how much product is allocated to each facility. Wadala typically manufactures product in large plants to lower production costs by exploiting economies of scale. Product is shipped to numerous customer locations. The design of its distribution network determines the total cost of delivering products to meet customer demand while maintaining the appropriate service levels.
There are many ways to configure a network to meet these goals. For example, a company can reduce its inventory holding cost by risk-pooling the inventory in a few warehouses. However, this option incurs higher transportation costs and longer lead times. Conversely, a company could expedite its response to demand by stocking inventory in a large number of warehouses. But such a strategy requires the company to maintain higher inventory volumes in order to provide the same level of product availability, which involves higher inventory carrying costs.
The thesis assessed the robustness of different distribution network designs.
First, an optimization model was designed to minimize the total cost, which included the costs associated with transporting product from plants to warehouses and from warehouses to customer locations, opening and closing warehouses in different locations, fixed warehouse operations, and maintaining inventory. The model also had to comply with various constraints such as limitations on warehouse periods of operation and the need to meet minimum safety stock levels. The key consideration was deciding how many warehouses the company should support, and which time periods the facilities should operate within.
Second, a list of critical uncertainties affecting the optimality of the distribution network was compiled by reviewing the literature and interviewing Wadala’s supply chain planners and managers. Guided by the variables in the optimization model, it was determined that four business and macroeconomic factors influenced Wadala’s network design decision: demand growth, oil prices, industrial real estate prices, and interest rates. The range for plausible values of the above mentioned environmental factors over Wadala’s planning horizon was determined using expert opinions gathered via desk research.
Third, multiple plausible scenarios were defined, based on combinations of the extreme values of these factors. The optimal network design was found for each scenario using the mathematical optimization model. Further, for each network design, the cost difference between each given scenario and its optimal version was calculated (known as the “regret” associated with each individual design).
Fourth, further analysis of the regrets of various designs under different scenarios revealed that for Wadala, the most important parameter for achieving drastic changes in the efficiency of a network design was the price of oil. A more detailed regret-based analysis considered small increments in the price of oil to identify the distribution network design that would minimize the total regret.
Three key observations can be drawn from the research.
Mathematical optimization can be enhanced using qualitative scenario thinking. The methodology developed in this study uses quantitative mathematical optimization for designing distribution networks, and enhances this approach by considering a broad range of variable factors using qualitative uncertainty analysis techniques borrowed from scenario planning. The combined optimization-scenario analysis approach helps highlight the critical environmental factor(s) that have the largest influence on the performance of a distribution network.
Network designers can use quantitative scenarios
The methodology helps a company to isolate and focus on the few major environmental factors affecting the robustness of a distribution network design. A range of plausible values of these factors, based on expert predictions, are used to estimate the performance of a network under various quantitative scenarios.
Robust design using regret-based approach
The methodology identifies a network design that minimizes regret for the company for different scenarios of the most important factor(s). While all the scenarios considered in the research were considered to be equally likely, it is possible to represent the likelihood of different scenarios occurring.
For more information on this research project and the MIT-Malaysia Master of Science in Supply Chain Management program, please contact Dr. Shardul Phadnis at email@example.com.