Rules for the Contest

This section is the most interesting one. The HTML version of the rule is somewhat ``garbled'' in equations. The handicapped time should have been $R+\frac{S}{30}-B$, in which R is the time taken for the run, S is the start time of the run, and B is 10 seconds if the robot has remained untouched up to the end of the run.

This equation favors autonomous robots that do not require any intervention. Also, note that a run time is ``the time it takes for a Micromouse to travel from the start square to the destination square.'' This means manually moving the robot from the destination back to the start square is worth it if it will take $10 \times 30$ or 5 minutes for the robot to autonomously travel back to the first square.

The time limit is 15 minutes. Because of the start time component in the handicapped time, later runs are generally less likely to take less handicapped time even if the paths are more optimized.

The following statement is a little value:

The starting procedure of the Micromouse shall not offer a choice of strategies to the handler.

Obviously, you cannot flip switches when you are string the robot for a competition. However, this does not mean you cannot have switches and alternative strategies. You just have to choose the strategies before the maze is unveiled.

In addition, the robot can autonomously choose the next strategy based on what it has learned.

The rules explicitly say that the illumination, temperature and humidity of the competition site can be changed ``at the discretion of the contest officials''. This means don't count on it. Make you robot so it can handle a tough environment (it'll still be in door and somewhat comfortable to people).

Note that if you replace parts of the robot, you may be requested to erase memory of the maze. This means don't let your robot run out of battery in the middle of a competition.

There is no rule regarding memory of the maze when a robot is manually placed to the start square because it ``appears to be malfunctioning''. ``Malfunctioning'' typically includes crashing into a wall. This means you can let the robot retain most of the map (except for the last cell or two right before it crashed).

In order to observe the rule that says no strategy offering at start up, you need to design the logic of the mouse such that it clears the maze once when it is powered up (which is not a choice). Ever since this powering up, the robot retains the map even when it crashes. You use one single button to instruct the robot to begin from the start square again after it is picked up. In fact, the robot should stop when the button is pushed, and only restart after the button is released. This gives the handler time to place it in the maze.

If you decide to retain the map after a crash, you have to be careful about the last few walls entered into the map. They can be wrong! If in doubt, erase more walls. All useful algorithms are greedy, which means even if the robot assumes fewer walls then there really are, it will still be able to explore and eventually figure out the shortest path.