Add multi-platform CMake workflow configuration#3
Open
thtran97 wants to merge 1 commit into
Open
Conversation
thtran97
pushed a commit
that referenced
this pull request
Jun 18, 2026
Re-enables conflict learning for one iteration after each restart. Controlled by KBILSSolverConfig::adaptive_learning (default false). New bench variant KB-F = KB-C + adaptive_learning for comparison. Result on Cordeau pr01-pr20 (10s, seed=42, 1 vehicle): KB-F rewards identical to KB-C on all 20 instances. KB-F accumulates 10-100x more clauses (68K vs 618 on pr01) with no quality gain — post-restart infeasible pairs are largely duplicates of conflicts learned in iterations 0-99. KB structural convergence is fast. Conclusion: adaptive_learning does not improve solution quality on this benchmark; KB-C (static, 100 iterations) remains the recommended config. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LhpxDkKFqtsxX9XMHPwVqy
thtran97
pushed a commit
that referenced
this pull request
Jun 18, 2026
RouteOrder: for a failed insertion (c, v), find the insertion position minimising detour cost in vehicle v's route, then take a window of (max_scope_size-1) consecutive route clients centred on that position as the xplainer scope (kb_ls.cpp:guess_conflict_scope). Revert adaptive_learning: remove the flag and just_restarted logic from KBILSSolverConfig / do_solve — the negative result from #3 is committed in the prior message; keeping a dead flag in the API adds no value. Evaluation on Cordeau pr01-pr20 (10s, iterations=0, seed=42, 1 vehicle): - KB-G (RouteOrder) generates ~0 KB clauses on 16/20 instances. Route neighbours are already feasible together, so no TW conflict exists within the chosen scope. NearestTW (picking clients with similar TW midpoints) is the right heuristic for conflict extraction. - KB-C = KB-G = ILS09 on all 20 instances under fair comparison. Earlier apparent gains (e.g. pr03: 377→387) were artefacts of different iteration budgets (1000 vs unlimited). Under matched budgets, ConflictStore variant #1 neither helps nor hurts on this benchmark. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LhpxDkKFqtsxX9XMHPwVqy
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.