The cost of an action is the weight of the box being pushed, this may be zero if there is no box or the box weight is zero, plus one for the worker’s movement.Īs mentioned earlier, the solution uses two search algorithms to find an optimal solution, a breadth first search algorithm -to find all the cells reachable by the player character-, and an A* graph search -used to determine the optimal path for the Sokoban puzzle using a heuristic. Thus, the weight of the box must be considered when computing the total cost of a path. In this variant of the game, each box is assigned an individual pushing cost. Although, you may not move through walls or push more than one box at a time boxes can only be pushed, not pulled.

The player can move up, down, left, or right, given there is no obstacle in the way. I've only played with it for a couple of days, but it at least demonstrates the way hexagonal Sokoban could work.Sokoban is a one-player game in which the player’s objective is to push all the boxes onto designated storage locations. Skinner (maker of many excellent Sokoban puzzles) I made a modified version of JokoSoko which I am calling HexoSoko. N E W! JanuHexagonal Sokoban! Based on an idea suggested by David W. Watch this space for links to some of them. There are many Sokoban web sites and many hundreds of puzzles available. I also intend to put together some more Sokoban links on this page. I will be submitting the Euphoria source to the archives if you have any interest in seeing it.įuture enhancements include mouse usage, puzzle editing, and a few other ideas I have in mind.

I developed the game using the Euphoria programming language. I'll probably add more puzzles later as I develop them. I've included six puzzles that I developed myself. It can record and play back macro sequences so that repetitive moves can be done quickly and easily, allowing more time for actually solving the puzzle instead of repeatedly stepping through sequences that have already been solved. It tracks up to ten scores for each solved puzzle and can replay the solutions at whatever speed you like. It allows you to undo moves back to the start of the puzzle. I still have some more work to do, but the game plays accurately for any size puzzle (at least, I've never seen one too big for it). But I have now put together the most complete implementation I've ever done: As I've gotten different PCs I've found different versions of the game and even written a few myself. I first discovered it on a Unix box at Georgia Tech sometime in the 1980's.

Online sokoban solver