Question:
have there been serious attempts to design divide and conquer heuristics for generating near optimal Hamilton Cycles in complete symmetric graphs?
For clarification:
by a divide and conquer heuristic for TSP I mean
- recursively partitioning the set of vertices with minimal size difference (i.e. equal for even size and difference 1 for odd sizes),
create at most two TSP from the partions until the instances can be optimally solved - repeatedly creating a graph with a small number of Hamilton Cycles, e.g. a ladder graph, from the tours through both parts and return the Hamilton Cycle with optimal length.