Abstract. This paper deals with self-organization phenomenon in qualitative microscopic pedestrian simulation. The agent-based pedestrian model in NetLogo is presented. Within the model, the lane formation is identified as the emerging pattern growing from counter flows of individuals. Information entropy is applied in analytical component of the model with the aim to measure the level of self-organization. Experimental results are provided.Keywords: Pedestrian simulation · Self-organization · Information entropy · Lane formation · Multi-agent systems · NetLogo IntroductionPedestrian and crowd motion models help us to identify and analyze spatial walking patterns under normal or competitive situations and to reproduce empirically observed crowd features. Moreover, similarly to models of fish schools or ant colonies, pedestrian simulations can serve as an experimental area for studying self-organization phenomena within computational methods. Being inspired by [3,4,6,7,9, 11] our idea was to apply information entropy to measure the level of self-organization. To achieve this we built a pedestrian model with analytical component for entropy calculation. According the level of abstraction, there are three classes of pedestrian and crowd models: microscopic models which describe each pedestrian as a unique entity, macroscopic models which aggregate pedestrian dynamics by flows or densities and mesoscopic models which operate with velocity distributions. Theoretically, pedestrian and crowd simulations can build on agent-based approach where individuals are seen as autonomous, rational, adaptive entities. In physics-based models such as social-force models or fluid dynamics models, pedestrians are seen as particles which are under the pressure of external attractive and repulsive forces. In cellular automata models, main concept is the walkable space which is represented by the lattice; its each cell is occupied by nobody or one pedestrian and in each time step pedestrians move to unoccupied neighboring cells. In queuing models the space is reduced into the network of nodes and links, with pedestrians moving around it, but without details of dynamics inside nodes. For further details on pedestrian modelling principles, see e.g. [1,2,5,8].
Rare attempts to use knowledge technologies and other relevant approaches are found in the river basin management. Some applications of expert systems as well as utilization of soft computing techniques (as neural networks or genetic algorithms) are known in an experimental level. Knowledge management approaches still have not been used at all. In this paper we discuss knowledge-based approaches in the river basin management as a difficult yet important direction which could be proven to be helpful. We summarize the research done in the scope of the AQUIN project, one of first Czech knowledge management projects in the river basin management. The project was initiated by the water management company in Pilsen, where dispatchers make decisions about manipulations on the reservoir Nýrsko, the strategic source of drinking water for inhabitants of Pilsen. The project aim was to support dispatchers' decision making under a high degree of uncertainty or data shortage. The research is continued in the scope of a new project AQUINpro, planned for the period of 2006 to 2008.
The objective of the chapter is to identify and analyze key aspects and possibilities of Ambient Intelligence (AmI) applications in educational processes and institutions (universities), as well as to present a couple of possible visions for these applications. A number of related problems are discussed as well, namely agent-based AmI application architectures. Results of a brief survey among optional users of these applications are presented as well.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.