A 512x512-pixel 8-bit image of a 2-dimensional array of simple interacting agents after 1000 iterations starting from a random distribution. Every agent has the same rules for determining its behavior. First the simple logistic equation is used to compute the next state from the current state: X(n+1) = BX(n)[1-X(n)]. Then an additional term describing it\rquote s communication with its environment - the average value of its 4 nearest neighbors - is added. The values of are clipped to constrain them between 0 and 1. Patterns emerge from the collective behavior of this collection of communicating agents [Abraham and Corliss, Jour. Bifurcations and Complexity, vol. 1, p 68-74, 1992].\par The computations were carried out on the original one-of-a-kind Massively Parallel Processor at NASA Goddard Space Flight Center as an NAS-NRC Senior Research Associate in the Computer Systems Research Facility headed effectively by Jim Fischer<>. The FORTH environment I used was created by my mentor in computation then and now, John Dorband <>.
Computer Science is no more about computers than astronomy
is about telescopes - E.W. Dijkstra (in Wired) |