Path optimization algorithm for long segments search and small segments minimization

Hi all,

I am developing a code to find the best subpaths in a complex graph. My graph is just a representation of a physical structure, so my edges are actual segments. When I say "best" I want to obtain the longest segments possible that can also minimize the resulting "short" segments (only one edge). What I am doing now is an iterative approach.

Given a maximum length of the sub-path:,

• each iteration starts from one of the n available nodes
• the n subpaths are generated and the best one selected. Each subpath is generated based on available neighbors (not completed nodes and not visited edges)
• I have a function to remove selected subpaths (and nodes with no available neighbors) from the next subpath generation search

I wanted to implement something metaheuristic like cuckoo nest, or bat algorithm, but my conditions here are slightly different. In my case the nodes must be visited more than once, because I need all the nodes to be included in the final count. This means that a node is excluded from the search when it's completed, not just visited.

My code is working fine, the problem is that I still have a high number of short segments (still 30% of the total number of edges, which is a lot), so perhaps I am missing another perspective or another approach that I have not thought about!

Thanks in advance for your help!