Anwar Haredy

Software Engieering and Computer Science

Vacuum Cleaner World


I was reading Artificial Intelligence: A Modern Approach, an introductory book about the field of AI and its fundamentals and reached a certain programming exercise which I then tried to solve.

The exercise

The exercise asked to design an environment in which a hypothetical smart vacuum cleaner operates. The environment is a simple grid, with each cell either containing dirt, obstacles or nothing. The vacuum cleaner must then roam the grid and clean the dirt as good as possible (with as little moves as possible cleaning all dirt etc).

Finally, the cleaner must return to its starting cell.

My attempt?

I struggled to come up with a solution that acts like a real AI (i.e makes decisions based on precepts). But I thought it would still be fun to program the environment and solve it in any way.

I realized that an (N * N) grid (or any grid really) can be thought of as a graph. Each cell is a node that is connected to its four adjacent nodes (top, down, left, right).

My solution

I implemented a simple depth-first search algorithm with the above realization in mind. And then for fun I visualized the solution on an HTML page using React, which was NOT the tool for the job but was fun nonetheless.



The UI is simple. The obstacle chance specifics the chance of a random cell containing a wardrobe. The dirt chance similarity does the same but for dirt and size is for grid size.

Clicking start will run the vacuum cleaner through the space.

Try it out hereeeeee!