> comments by WDS > > 1. I think using utility=-distance > is not as realistic as something like > utility=1/sqrt(1+distance^2) > > I claim the latter is more realistic both near 0 distance and near > infinite distance. > > 2. It has been argued that L2 distance may not be as realistic > as L1 distance. > L2=euclidean > L1=taxicab >
Suppose that the "candidates" (i.e.alternatives) are possible locations for a building, and that the inconvenience of each alternative for each voter is proportional to the distance from that voter's residence to the location, or simply the time it takes to get there. The taxicab distance would be a natural metric in this situation, but I don't see utility = 1/(1 + distance^2). If I were a voter in this situation, my sincere rating for an alternative at distance x would be r=(D-x)/(D-d), where D and d, respectively, are the distances to the respective alternatives furthest and nearest to me. I don't see how that could come from normalizing the suggested utility of u = 1/(1+x^2) . Forest ---- Election-Methods mailing list - see http://electorama.com/em for list info
