An Integrated Production , Distribution

The critical importance of integrating production, distribution, and inventory (POI) operations has long been recognized by top management of many companies. Now. using the latest advances in Management Science modeling and solution technology, an integrated computer-based POI system has saved approximately $18 million dollars during its first three years of implementation for a major national firm. Agrico Chemical Company. According to the Vice-President of Agrico Supply and Distribution, an additional $25 million savings is anticipated over the next two years. Brought about by close cooperation between company officials and an outside staff of Management Science consultants. .the POI system has been used extensively to evaluate the benefit/cost impact of alternative capital investments in both short-term and long-term planning decisions. The development of the system underscores the value of recent Management Science innovations that have made it possible to analyze interacting influences too numerous and complex to be analyzed adequately only a few years ago. Advanced network methodology incorporated into the POI system required only one one-hundredth of the computer time and coSt of methodologies previously used. The power and flexibility of the new Management Science tools have also brought about increased communication and understanding of key company operations. This increased communication and understanding stems from the inherent .'pictorial.' nature of network-based models. which facilitates interpretation of these models and policy recommendations based upon their solution. . IN VENTOR Y IPRODUCTION: NETWORKS/GRAPHS INTERFACEJ N,'\'eml1er 1979 21 Overview Agrico Chemical Company, with annual sales exceeding half a billion dollars, is one of the nation's largest chemical fertilizer companies. A subsidiary of The Williams Companies, Agrico mines, manufactures, and markets eight principal chemical products domestically and internationally. The company's success story, based on aggressive and forward-looking management, is typical of others in which a relatively small firm has been transformed into a leader in its field in less than a decade. In the mid-1970's Agrico encountered unexpected difficulties. The seasonal demand characteristic of the chemical fertilizer industry was creating a chain of intricate and far reaching effects that could not be responded to adequately. As a result, the company's profit margins were being seriously eroded by steeply escalating distribution costs. It became apparent that a multitude of interdependent factors made it impossible to find a remedy through customary methods, such as studying cost figures and charts. The web of interacting ihfluencf:s which spanned the company's principal activities -production, distribution, and inventory -required an integrated computer-based planning system to uncover the appropriate decisions. In 1976, David Wilson, Vice-President of Agrico Supply and Distribution, in coordination with Herb Beattie, Vice-President of The Williams Companies Information Services, created a project team to develop such a planning system. The principal objective of this team, headed by the authors, was to develop a computer-based production, distribution, and inventory (PDI) planning system which integrated the three major segments of Agrico's business decisions: (1) the Supply Segment, consisting of production, purchases, and product exchanges with other chemical corporations (coproducers); (2) the Storage and Customer Distribution Segment, involving sizing and locating bulk distribution centers; and (3) the Demand Segment, involving customer demand throughout the eastern two-thirds of the United States, and locations where the product must be supplied to coproducers due to product exchange agreements. The project team designed and implemented a POI system utilizing recent advances in network modeling and solution technology [I], [2], [5-10], [12], [1416]. The system was given the capability to provide planners with insight into the system-wide ramifications of their decisions. It~ integrated framework allows the system to consider the relevant environmental impacts of all decisions simultaneously, thereby equipping it to provide analyses for long.. and short-range planning and operational decisions. Long-range planning summary In long-range planning, the system is used primarily for decisions associated with the sizing and configuration of the distribution system. This helps to answer such questions as: -Where should distribution centers be located and what should be their size? -How much long-term inventory investment should be made? INTERFACES November 1979 22 -How much transportation equipment is needed? -What supply/purchase/exchange opportunities should be exploited? In the short time that the integrated network POI planning system has been in use at Agrico. it has already proven to be an extremely valuable decision aid for long-range planning and its effect in cutting the steeply rising distribution costs has been dramatic. Further. the system has uncovered several entirely unanticipated areas of cost savings. primarily through the evaluation of capital investment decisions. One of Agrico's long-range studies, using this system, showed that by locating a distribution center on the upper Ohio River, $100,000 could be saved on transportation costs. Another study revealed that it was worth $175,000 to obtain the early completion of a new distribution center. The PDI planning system was also used to evaluate the long-range impact of changing the distribution pattern of a particular nitrogen chemical solution plant. Prior to the analysis using the PDI system, Agrico management had planned to build an additional 100,000 square foot storage facility to accommodate the forecasted growth in demand over a five-year period. The PDI planning system revealed that the additional storage facility was not cost effective. By not expanding the storage capacity Agrico realized a capital savings of $800,000. In addition, the distribution plan suggested by the POI system reduced the annual transportation costs for the plant by $12,000. Other long-range planning studies, using the PDI system, have been carried out by Agrico management, but the financial impact of the resulting decisions is more difficult to quantify. One such decision involved the closing of three small distribution centers in the midwest. The resulting capital was reinvested in rolling transportation equipment. In light of the current freight car shortage and energy cost increases, it is virtually impossible to accurately gauge the long-range affect of this decision. The total quantifiable impact of Agrico's long-range planning studies using the PDI system is a cost savings in excess of $1 million. However, the most substantial results have come from the usage of the PDI system for short-range planning and operati()nal decisions. Short-range planning summary For short-range operational decisions, Agrico uses the integrated network-based PDI system to aid in answering questions dealing with the allocation of a defined supply of product through a specific distribution center configuration. For such decisions, the system is used to avoid unnecessary production, distribution, and inventory costs by providing the capability to evaluate alternatives on a systemwide basis. To illustrate, the model is used to decide what, where, and how much product should be produced as well as when, where, and how much product should be shipped. Since 1976, Agrico has used the POI planning system for operational decisions to supply its products to customers at the right time and at the least possible cost consistent with good customer service. During its first year of implementation, the POI planning system enabled management to reduce Agrico's total distribution costs by $3.7 million. But this is not the bottom line. 23 INTERFACES November 1979 During the first year, Agrico was unable to adopt completely the decision alternatives suggested by the POI planning system. This partial implementation was anticipated, and is natural, since the task of rescheduling production and distribution activities is formidable. The overall impact of the planning system is best illustrated by Figure 1. This figure shows the growth in the total cost of distribution over the past few years as well as the anticipated growth through 1980. In the figure, the projected distribution costs are those that Agrico would have expected to incur if the POI planning system had not been developed. The projected distribution costs are based on a conservative 7% annual inflation rate applied to the known point-to-point freight rates prior to the development of the POI system. These inflated freight rates are multiplied by the anticipated distribution tonnage to obtain the projected costs. The actual incurred distribution costs through ..978 are also shown in Figure I. These are projected forward for 1979 and 1980 by applying the same 7% annual inflation rate of the 1978 point-to-point freight rates obtained by the POI system. A total distribution cost savings of nearly $17 million was made from 1976 to 1978. It is projected that by 1980 the POI planning system will be providing Agrico with an annual distribution cost savings in excess of $13 million. Clearly the impact of this planning system is becoming more pronounced as the cost of energy increases. FIGURE Distribution Cost Trends