2006 Fiscal Year Final Research Report Summary
Study on Traffic Simulation and Cognitive Bias
| Project/Area Number |
17500142
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Sensitivity informatics/Soft computing
|
|---|
| Research Institution | Nagoya University |
|---|
Principal Investigator |
KITA Eisuke Nagoya University, Graduate School of Information Sciences, Associate Professor, 情報科学研究科, 助教授 (50234224)
|
|---|
| Project Period (FY) |
2005 – 2006
|
| Keywords | Traffic Simulation / Stochastic Velocity Model / Traffic Survey |
|---|
| Research Abstract |
The object of this study is the basic study on the traffic simulation concerning driver's cognitive bias. The research project can be classified into two groups; one group investigates real traffic data to determine the relationship between the vehicle behavior parameters, and another group constructs mathematical model of human cognitive bias.
For determining the relationships related to the vehicle velocity and distance, we took actual traffic flow on a video. One determined from the video, the relationships between a velocity and vehicles distance and a velocity and a acceleration ratio. It is concerned that each vehicles drive according to the relationships and then, two traffic flow simulations on a two-lane freeway were performed. In one simulation, all vehicles move according to same relationships and, in another simulation, they move according to individual relationships. The results show that the maximum traffic amount in the latter case was smaller than that in the former case.
For modeling the cognitive bias, the stock investment in the artificial market was considered. The artificial market was constructed as the multi-agents system. The cognitive bias of the agents are modeled according to the prospect theory. The stock fluctuation of ten companies was simulated by the efficient agents and the agents according to the cognitive bias. The results show that the agents according to cognitive bias can simulate the stock fluctuations better.
|
