Abstract
Opinion distributions in a society consisting of 1, 600 people were simulated using a cell-automaton model. In the simulation, each person was assumed to hold one of two opinions at a certain point of time, and to change his/her opinion according to an assigned probability function. One of three types of probability functions, representing three types of long-term attitudes, (i. e. supporting the first of the two opinions, supporting the second of the two opinions, and neutral) was assigned to each person, using a predetermined ratio. The results showed that, when the ratio of the three attitudes were varied, a society was likely to face one of the following three situations in the near future: (1) a single type of opinion distribution was expected to be attained, (2) two types of opinion distributions were expected depending on the initial situation, and (3) every possible opinion distribution was expected with almost the same probability.
