It seems that the A* algorithm is not delivering the performance I was hoping for, so I am trying a new approach. Are there any existing cases, by any chance? If not, I would like to hear your thoughts on this approach.