From the Shadows: Earthmoving on Construction Projects

The finalists' papers from the 1982 Management Science Achievement Award competition were the stars of December's special issue. Here, I commend to your attention an account of an LP application from Australia. It was not a finalist in the competition; indeed, it was not even submitted to the competition. But, by comparing it to the finalists' papers, we can draw some useful inferences about the potential value of solid LP applications.

The article deals with minimizing earth-moving costs at an airport construction project. The two resulting LP models are straightforward, simple, and small. The smaller one is a pure transportation model with 4 sources and 5 destinations, yielding an LP structure with 9 rows and 20 columns. The larger one is about twice as large in each dimension. In terms of model size and structure, this application might well be an illustrative example in an introductory MS/OR textbook. But it's not a made-up example; it's real. This tiny example produced demonstrable savings of about $400,000 over the best that experienced “engineering judgment” could yield and, consequently, influenced operations in the field run by non-LP-oriented managers.