Lessons Learned Delivering Optimized Supply Chain Planning to the Business World

time for business enterprise software in general, and software for optimized supply-chain planning in particular. The spread of enterprise resource-planning software, warehouse-management systems, factory-planning systems, and other enterprise applications, had the side effect of moving online much of the raw data needed to optimize supply chains. Furthermore, the pace of business change, increased competitive pressures to react quickly to this change, and the rapid development of online commerce forced businesses to question the week-plus supply-chain planning cycles that had been the norm. Finally, the year 2000 (Y2K) problem caused an across-the-board replacement of enterprise software, allowing many businesses to update their approach to supply-chain planning. The end result of all of these factors was a huge upswing in demand for supplychain planning tools from i2 Technologies and other vendors. When I joined i2 in 1996 as optimization architect, the company had around 250 employees and roughly $100 million in annual revenues. At its peak in around 1999, i2 had grown to have a market capitalization of over $13 billion, had over 600 customers, and was listed on the Nasdaq 100. This rate of growth brought into sharp relief two critical technical challenges in supply-chain management. First, the data required to manage supply chains was indeed online, but it was inevitably inaccurate and incomplete. Further, it resided in a host of incompatible and disconnected databases. Second, and equally important, supply-chain planning and optimization were a perpetual challenge. There are solid theoretical reasons why supply-chain planning and optimization are hard. Technically the underlying optimization problem is either NP or P-space complete (depending on the details of the domain). Furthermore, the problem mixes a dozen or so classic optimization problems from AI and operations research (OR), and much of the expected savings from global supply-chain optimization are lost if these problems are treated independently. This article describes our experience from four years of solving supply-chain planning and optimization problems across industries, and some of the lessons we learned.