"Net driving time" from two separate ambulance depots to local homes is marked by concentric rings of 0-2.5 minutes, 2.5 to 5 minutes, 5-7.5 minutes, 7.5 to 10 minutes, etc. Published several years ago by Ritsema Van Eck, it identifies areas not covered by a Dutch ambulance service on an island in the southwest of Holland. Illustration 1 presents a basic GIS approach to this class of problem. In effect, what is required is a methodology that maximizes a spatial relationship between potential supply sites and potential demand points across a variable surface whose features include population density, housing density and cost.
The solution requires consideration of a series of constraints involving cost, number of vehicles, distance measured in time and usage. "Optimization models provide coverage in nearness or relationship to all of the elements or members of a group." A classic problem is the siting of ambulances or firehouses in a manner that maximizes coverage across a region while minimizing the cost of that coverage. No special claims are made for either ArcView or Maptitude author familiarity with both programs influenced these examples.
While examples here rely primarily on two different GIS software packages ESRI's ArcView and Caliper's Maptitude there are an ever-growing number of GIS programs on the market today or coming to market. The intent is to provide a general survey for OR/MS readers, irrespective of their familiarity with GIS approaches. In coming years GIS programs may well become the engines that drive the analytics of health care planning, transportation modeling, locational analysis and political prognostics.Īs an introduction to these tools and their capabilities, a series of OR/MS-style problems ranging from the routine to the innovative are described. This is the promise of GIS: a single tool that permits the extraction and presentation of useful information from huge amounts of data. "The handling and processing of enormous amounts of data, the extraction of useful information, the fast processing of information, the decision-making and subsequent implementation of decisions and control inputs are research tasks that need to be addressed." They simultaneously present a problem definition and then evidence of its solution. "Whereas with conventional mapping (and even the earliest digital mapping) the map was the database, today the map is merely an evanescent projection of a particular view of a spatial database at a given time." Maps become, in GIS, synchronous essays displaying robust solutions to complex problems. Its utility lies in the ability to analyze spatially related data and then present it in a graphical form map, histogram, chart, or as a differentiated spatial surface that may or may not look like a standard graphical representation of the world.
In the following two decades the growing power of computers permitted more intense analysis with DOS-based programs for those who understood the essentials of mapping and of computer programming.Ī Windows environment, better data storage systems and a new generation of point-and-click database programs have transformed GIS into a general modeling tool. While fascinating to cartographers, it was of only esoteric interest to anyone else. In the 1960s, they used punch cards to enter the coordinates for locations so that automated plotters could draw a precise map of a specific region. Their power lies in the marriage of easily understandable graphics and increasingly powerful systems for data analysis. What they share is a means of graphically presenting (through maps and charts) data stored and manipulated in a related spreadsheet or database.
Within five years, they may subsume them.Īt present, there are a range of GIS programs all sharing similar attributes, each with its own strength and weakness. Today they complement a range of more traditional OR/MS tools. These technologies have become a primary environment for problem solution and general modeling. What began as a technique for automated cartography is maturing.
Its general shape is that of the current generation of geographic information systems (GIS) software. In 10 years, perhaps less, it will be mature. Then crunch the numbers in a manner that yields a solution page whose results are not only robust but also intuitively comprehensive, graphic and clear.Ī black box would be nice, wouldn't it? It would be a general tool whose utility ranged across the myriad issues defining the OR/MS problem set. In fact, throw in all the data variables you want. Meld time and distance as equal variables. Combine the perspectives of locational analysis and the general constructs of social theory. Marry linear programming to simulation modeling.