4.2.3 Choice of the weighting parameters

A random attitude of the weighting parameters forbids itself because you do not get sensible result.

If one chooses weighting for markers $\alpha$ too small, the cooperation ability between the ants is limited much. You'll receive a small search room per agent and the procedure resembles a GREEDY Algorithm.

If one chooses weighting for markers $\alpha$ too large, it leads to the over interpretation of the marker thicknesses of other ants: all ants follow the same tour and it comes to stagnation.

If one chooses weighting for markers $\beta$ too small (less than 1),the algorithm works independent of the distances between the knots, but the TS problem and/or the algorithm depend on it!

The weight of the inverted length $\beta$ should be between 1 and 5, in dependence of it weighting for markers $\alpha$ between 0.5 and 1. With these settings we were able to find the best (or close to it) solution of several TS problems. Moreover, the evaporation factor can still be chosen between [0,1]. Where 1 means no evaporation, and 0 means total evaporation. We were able to make good experiences with values between 0.5 and 1.