An intelligent decision support system for flood management: A spatial system dynamics approach.

摘要

Natural and human systems are dynamic and complex. Dynamic complexity arises because systems are: non linear; governed by feedback; adaptive; and counterintuitive. When modeling social or physical/natural processes, feedback occurs between time and space and a system's performance in time is influenced by the change in conditions in space and vice versa. The current approaches being used for dynamic modeling either ignore the feedbacks or do not explicitly consider spatial-temporal interactions.; A new approach called spatial system dynamics (SSD) is developed to model feedback based dynamic processes in time and space. This approach is grounded in control theory for distributed parameter systems. System dynamics (SD) and geographic information systems (GIS) are coupled to develop this modeling approach. A mathematical formulation is presented for one and two dimensional distributed parameter control problems along with their finite difference solutions. The SSD approach is capable of handling both, real and Newtonian time. It can integrate different models (hydrologic, hydrodynamic, economic, and behavioral) in a single modeling environment. These models run simultaneously and communicate with each other by providing feedback.; The SSD approach is superior to existing techniques for dynamic modeling, such as cellular automata and GIS, and it addresses most of the limitations present in these other approaches. The SSD approach can be used to model a variety of physical and natural processes where the main interest concerns the ‘space-time’ interaction, e.g., climate change; disaster management (spread of infectious diseases, fire spread, overland flooding); environmental processes; and natural resources management.; The spatial system dynamics approach is used to develop an intelligent decision support system (IDSS) for flood management. The IDSS is able to assist in: forecasting floods (using artificial neural networks); selecting suitable flood damage reduction options (using an expert system approach); modeling the operation of flood control structures; describing the impacts (area flooded and damage) of floods in time and space; performing economic analysis; and planning evacuation. The IDSS is implemented for the Red River Basin in Manitoba, Canada. The IDSS environment allows a number of what-if questions to be asked and answered, thus, multiple decisions can be tried without having to deal with the consequences.