Is One Route Good Enough? Continued from page 33 ations, and construction time. Once the pipe network is identified and the decision matrix is populated with the relative scores and weights for all the segments and criteria, the application is ready for the analysis. The analysis process consists of three steps: 1. Import segment weights and costs into ArcGIS. 2. Identify potential corridors. 3. Export results back to Microsoft Excel. Optimizing a corridor from a nest of segments is similar to finding the directions through a network of roads—at least one attribute is required for the optimization process. Minimizing an objective function is the key to determining the best route. In this analysis, two attributes are used: total segment weight and cost. Once the decision matrix is populated with criteria weights and segment costs and scores for all criteria, it is important to import the segment weight and cost into the Pipe Segments feature class in ArcGIS. From ArcGIS, the application lets users browse for the Microsoft Excel decision matrix. Once the file is opened, the application looks up the segment weights and costs for each segment based on segment ID in the segments feature class and reads the data from Microsoft Excel cells, which contain weight and cost. The application highlights each cell in the decision matrix to indicate is has been successfully read. The application then appends three fields to the Pipe Segments attribute table. The SgLength field contains the length of each segment in map units, which is feet. The SgWeight field contains segment weights imported from the decision matrix. The SgCost field contains segment costs imported from the decision matrix. However, if these fields exist before the application runs, the values in these fields are just updated. Once the individual segments', weights and costs are imported from the decision matrix, the application is ready to perform the Network Analyst process. The application loops through the segment feature class and identifies all potential routes from the start node to the end node(s). The algorithm for this command minimizes the total weight of the segments comprising the current corridor. It does not take the cost of each segment into consideration in selecting the corridor but does keep track of the cost and sums it for each potential corridor. After successfully identifying routes, the application generates a new feature class and Test case pipe network feature class with 79 segments, one origin point (point 1), and two destination points (points 2 and 3) An example of the best of the top 31 routes selected in an application run 34 ArcUser Spring 2010 www.esri.com