# Place points to minimize population weighted euclidean distance to census centroids?

In ArcGIS 10.22: I need to take about 3000 contiguous census blocks and determine where to place 15 points such that the points minimize the population-weighted Euclidean distance to the census block centroids. I'm trying to think of several ways to do it, such as creating a surface using population as the z-value. What would be the most efficient way to calculate this? Which tools would you use? (think, I'm going to drop x piles of corn in the area of varying density such that it would minimize the flying distance for the most crows to fly to those corn piles)

• Simple splitting them into 15 groups is challenge. What you are after is ginormous even for low resolution solution. Say you place 100 regular points and try their combinations by 15. Number of iterations is going to be 2.53338471349989E+17 (!) It might worth trying to play with groups of equal population, see this gis.stackexchange.com/questions/153094/… – FelixIP Aug 12 '16 at 6:35
• I kind of was figuring that, but assuming I could make the data more coarse, what approach(es) might be used? (aggregating the data could be part of the answer) – Nick D'Amato Aug 12 '16 at 14:55