collect statistics about optimal predecessor vector size depending on exploration depth
priority queue for expanding using the rook-distance / min distance between target robot and goal.
implement color-agnostic variant for GUI. Need to convert color-agnostic paths into colored paths.
try to overcome big std::map for insert and search time...
is it possible to write all reached states of one expansion level into a vector, sort it afterwards?
map insert is not a good operation when looking at allocation.
passing a custom allocator can be a solution
destruction of big containers has to be done in a separate thread. when solver data is thrown away (switching to a new game e.g.) dtors should not block UI
try to use a hash-map
what about std::make_heap on a std::vector to make deallocation more performant?
Ideas of further engine improvement:
collect statistics about optimal predecessor vector size depending on exploration depth
priority queue for expanding using the rook-distance / min distance between target robot and goal.
implement color-agnostic variant for GUI. Need to convert color-agnostic paths into colored paths.
try to overcome big std::map for insert and search time...
destruction of big containers has to be done in a separate thread. when solver data is thrown away (switching to a new game e.g.) dtors should not block UI
try to use a hash-map
what about
std::make_heap
on astd::vector
to make deallocation more performant?