For several years, we have been developing a general purpose road-traffic simulation system to analyze road traffic jam. This paper describes the concept of the system using the running line model, and a case study for general purpose simulation with cell aotpmaton model. In order to simulate congestion of road traffic system, it is indispensable to describe vehicles having their own decision-making capabilities, and to have detailed and exact road condition data on the road system. The road system is composed of such roads as streets, avenues and highways, and their intersections. In this study, we represent it by a network in which roads are arcs and intersections are nodes. The road-network itself has two kinds of attributes; the physical attributes and the logical attributes. The former are ones of the roads themselves such as their varying widths, shapes, slopes, running lines of traffic lanes and their channels. The latter are ones which depends traffic regulations such as signals and traffic signs. The former is called road information and the latter is road traffic regulation information in the road-network. Almost all road traffic simulations using a microscopic model have been carried out through a locally fixed road-network information database system. However, these systems are not effective. A general-purpose road traffic database system is therefore desired to analyze traffic jams in all areas. Because such a general-purpose system is very voluminous, speedy input of road network attributes should be taken into consideration, and the addition of some effective error-correcting method. We propose a new concept whereby the running lines on the various shapes of the road can be represented approximately by a combination of lines, circles and clothoid curves allowing speedy calculation. We developed a general-purpose database system to describe road information except for the road traffic regulation information. It is important to draw quickly and easily any running lines on the road of the database system. On the other hand, several studies have been done to realize road traffic simulation by a microscopic model. For example, there is a flattery model of a vehicle by a fuzzy theory and various theories such as a neural network work. A model could do movement every one vehicle in detail by using these methods, and high simulation of precision was enabled. However, these models are very complicated, and simulation practice processing time becomes long. Therefore, as a method we simplify a complicated phenomenon as much as possible, and to express, simulation by a cell automaton attracts attention. A cell automaton assigns one individual to one cell basically and moves a cell in consideration of being affected by the individual which each individual is next to. Modeling for simulation is clear and becomes simple a generous thing called a cell follows a specific rule, and to move. It may be thought that an error grows big when we cut off a small error. However, you should have you just think like the middle of a micro-model and a macro-model. We can realize total movement in form to be almost reality to some extent by doing simulation of a rough model. A cell automaton is applied in a wide field by such a characteristic. By this study, we changed passing and a traffic lane of it more, and parking vehicle evasion wrestled for realization of the run model that became important by a town area run. We introduce a model general idea to call "a temporary lane" for these run models. And these modeling, we can realize simulation to be almost reality. Because we did systems construction by these theories and the inspection by this thesis, we introduce about it.
