Added two sets of problem instances based on TSPLIB: homogeneous and heterogeneous
Implemented missing EvaluateByCoordinates for 1-shift moves
Made it clear that move evaluators are for estimated PTSP only
Changed parameter realization from a rather strange list of list of ints to a list of bool arrays
Reusing code of the 2-opt and 1-shift move evaluators in 2.5 move evaluator
Introducing distance calculators to properly handle the case when no distance matrix is given (previous code only worked with distance matrix and without only with euclidean distance in some settings)