Effects of Information Gathering on Imitation Behavior in One- Dimensional Cellular Automata:Computer Simulations Based on Prisoner''s Dilemma Moedls

• Main Authors: Chen, Chieh-Yuan, 陳建元
• Other Authors: Lai Shin-Kung
• Format: Others
• Language: zh-TW
• Published: 1996
• Online Access: http://ndltd.ncl.edu.tw/handle/02940816401374962743

碩士 === 國立中興大學 === 都市計劃研究所 === 84 === Because most of the research on planning behavior dose not consider the spatial element, this thesis explorse effects of information gathering in planning behavior on evolution of spatial systems. Because the prisoner''s dilemma game incorporates simple group interactions and can be used to simulate land development behavior, the research uses that model as decision rules in spatial interactions. However, the prisoner''s dilemma game dose not consider the spatial element either;therefore,the simulations conconducted in the research are based on models of cellular automata which are spatial in nature and have been thoroughly studied,in combination with the prisoner''s dilemma game and planning elements for computer simulations. From these simple spatial interaction models, we can observe how information gathering affects structural features of evolutions of these systems. It was found that increasee in planning investmentresulted in cellular automata more sensitive to the payoff parameter when the interactionrange was small, while less sensitive when that range became large, but the difference was not significant. Increase in planning investment also resulted in more diverse, complex evolution patterns. When the interaction range increased holding planning investment constant,the range of the payoff parameter was reduced and cellular automata were more sensitive to the payoff parameter. That is, small changes in the payoff parameter led to change in evolution rules. In addition, the C (cooperate)group in whichagents were adjacent to each other was in a better position than the D (defect)group to survive, implying that the spatial element had effects on system evolutions and that when applying economic decision models in spatial models, the equilibrium of the prisoner''s dilemma game need be modified. According to Wolfram''s classification, the cellular automata resulting from the computer simulations falled into class one and two, with only one exception belonging to class three or four. Most results from the simulations evolved into stable, predictable structures. When the range of cells'' values, The interaction range, and planning investment increase,it is more likely that cellular automata would fall into class four. This implies that real world complex systems, such as cities, might have inherently class four structures, which can serve as a hypothesis for establishing urban growth models based on cellular atuomata. Key words:planning behavior, infornation gathing, prisoner''s dilemma,cellular automata,planning investment.