A Decision Support System for Corporate Planning in the New Zealand Dairy Industry
<p>Decisions; Support; Systems; Planning; Optimization; Heuristics; Network programming; Fixed Costs; Travelling salesman; Distances; Computer graphics. A Decision Support System (DSS) is described, the prime objective of which is to aid in the location of new investments in a multi-site, multi-product dairy processing company. A network program model is described which optimises the collection of milk from farm groups (netcells) and the allocation of the milk to a range of final products and byproducts through consideration of product prices, Process costs and transport costs. Constraints include process capacities, overtime capacities, and final product demands. Site dependant product yields are considered through use of an iteration procedure surrounding the network model. This procedure updates estimates of the mean company yield used to set upper and lower arc constraints in the product demand phase of the network model. Milk tanker collection distances are estimated by an expected travelling salesman distance method in conjunction with accurately measured netcell to factory 'bridging distances' and an inter-factory diversion network of road distances. To cope with daily fixed coat charges, a heuristic Procedure employing cost relaxations and a number of Pre-solution feasibility tests is used. Seasonally varying factors (milk supply, product yield and farms visited per tanker trip) are accommodated by solving the network model for the average day in each month for twelve months, then summing the results multiplied by the number of production days in each month. Implementation as a DSS was facilitated through use of an interactive computer system incorporating computer-generated graphic displays. Applications of the DSS to location planning, industry rationalization and other corporate planning activities are described. Recommendations on the use of the model to identify the feasible set of candidates for location studies are made, and methods for identifying the appropriate timing of investments are considered.</p>
