A potentially interesting randomisations is where pairs of groups are selected based on some neighbourhood condition. i.e. select group 1, then select group 2 from the set of groups within a specified neighbourhood of group 1.
A bonus option is to then weight by proximity.
This could be done for the curveball alg first, but potentially also the independent swaps algorithms.
A potentially interesting randomisations is where pairs of groups are selected based on some neighbourhood condition. i.e. select group 1, then select group 2 from the set of groups within a specified neighbourhood of group 1.
A bonus option is to then weight by proximity.
This could be done for the curveball alg first, but potentially also the independent swaps algorithms.