Altsolver#16

Open

tbvanderwoude wants to merge 17 commits into

benchmark-verificationfrom

Commits on Nov 23, 2025

Added straightforward implementation of ALT algorithm in ALTSolver
tbvanderwoude
committed
Much improved ALTSolver speed by switching from vec of grids to grid of (small-) vecs
tbvanderwoude
committed
Added benchmark of ALT, using primitive monte carlo generation of landmarks
tbvanderwoude
committed
Introduced generic dao_bench_solver, reusing much of the benchmarking logic for the four implementors of GridSolver
tbvanderwoude
committed
Changed successors, heuristic and success in astar_jps from FnMut to Fn as in practice they have always been Fn
tbvanderwoude
committed
Implemented greedy landmark selection algorithm from single initial landmark
tbvanderwoude
committed
Simplified approximate goal condition using D instead of repeated EQUAL_EDGE_COST if
tbvanderwoude
committed

Commits on Nov 25, 2025

Improved tightness of heuristic by adding missing base heuristic of straight line distance
tbvanderwoude
committed

Commits on Nov 26, 2025

Removed grid reference parameter from heuristic and cost, simplifying solver trait and corresponding solvers
tbvanderwoude
committed

Commits on Nov 30, 2025

Commits on Jan 3, 2026

Made new_minmax_greedy handle the case where no new landmark is found
tbvanderwoude
committed