evokes human decision making

Requires a Java-enabled browser (for example, Microsoft Internet Explorer 5.0 or later). The applet requires Java 1.4.1 or higher to run. It will not run on Windows 95 or Mac OS 8 or 9. On a 33K dialup connection, the Java applet will likely require two minutes or more to load. (Once it is loaded you may be able to run it offline.)

Load the model (Java applet) in a new browser window

Created with NetLogo 2.0, a multi-agent simulation program that can be programmed using the Logo language. NetLogo users: view/download model file (23K): maze.nlogo.


This model demonstrates a variety of maze-running strategies as they encounter a variety of randomly generated mazes. These strategies evoke and resemble human decision-making frameworks.

For simplicity, the mazes and the strategies are orthogonal -- the moves are at right angles, not diagonal. You can edit the mazes that are generated before your run them, using the DRAW PASSAGE and DRAW WALL buttons and using the mouse.

This model differs from real life in two ways:

1. We're asking you to run cheeseless mazes. Drawing on our life experiences, it is easy enough to imagine finding -- or missing -- the cheese. There are no monsters or ghosts, except in the Past Experience mode.

2. You can tell how you are doing. THE MONITOR (on the upper right) shows the number of unvisited squares. It's not ambiguous, and it doesn't lie.

The maze evokes life's possibilities and potentials. How well does your decision strategy or framework fulfill life's potential? How about that of your family, your organization, or your various governments?


1. EXPLORE AT RANDOM. This strategy is a common one in human affairs. The maze runner proceeds ahead until she arrives at a junction or turn, at which point a random path is taken.

WALL-FOLLOWER. This resembles an ideological or faith-based approach. The maze runner keeps a wall on his right (or left) at all times. You can stop and reverse the preference at any time using the "right-left?" switch. Wall-following works well in perfect mazes without chambers or loops. Try a non-perfect maze and see if you can explore all possibilities with a "pendulum" strategy, alternating right and left.

PAST EXPERIENCE. This strategy resembles the common human decision framework based on past experience, peer pressure, or on the stories we keep telling ourselves about our failures and successes (our beliefs). Some adaptation can occur: each time this strategy is run, a different theory of the maze is attempted. The purple ghosts you see represent the beliefs or fears of the maze-runner, and are not features of the maze itself.

EXPERT OPINION. This strategy resembles much expert opinion in that there is a definite method for seeking untrod passageways, as well as returning to the starting point. However, there is little effective monitoring, and some passages remain unseen.

MONITOR AS YOU GO. This maze-running strategy, an implementation of Tremaux's algorithm or rule, marks the passages that have been traversed, thus preventing the maze runner from traversing a passage twice in the same direction. Though it is effective and efficient for exhausting the possibilities of a fixed maze, if you stop the procedure during its run (click the black button) and add new passageways to what has already been run, the maze runner will not always find them.

FLOOD-FILL. This strategy resembles a paradigm shift or phase transition. It resembles decision making that simultaneously considers ecological, economic, and social factors for the short and long term, and includes everyone.


This model may be freely distributed, shared, and modified, provided sources are credited. Please credit James Steiner and the Managing Wholes simulations page:

The maze-creation portion of this model is based on a model by James P. Steiner.

This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike License. To view a copy of this license, visit
or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.

Updated 3 April 2004