| The rapid expansion of urban areas creates highly centralization of population and property. Such expansion brings opportunity and challenge together. Urban safety management is now facing with unprecedented challenge. Making full use of the emergency responding forces and ensuring the safety and stability of city operation has become the key problem in urban management today. In the real scenario, several emergency responding teams are dispatching to appropriate number of emergency calls. In traditional processing mechanism, such scheduling is carried out artificially and regardless of environment information (traffic, precipitation and wind speed). What’s more, the distribution of emergency responding forces is never considered spatially and temporally. In some proposed optimization algorithm, the problem of real-time emergency calls responding is solved as a special freight problem characterized by noneconomic, random-call and real-time. However, algorithm, considering both real-time emergency calls and global emergency responding forces distribution, usually takes too long for real-time system and especially for a single emergency call. Thus, this article proposes a bi-layer based emergency response mechanism. The separation of real-time emergency calls and global optimization remarkably reduce the delay of dispatching and decision-making for a single emergency call. Real-time calls are handled by a multi-agent system. Although the global view is deficiency, such mechanism is still able to provide a reliable solution. In addition to this, a global optimization process is used to optimize locations of idle team and fixes the sub-optimal solution created by the multi-agent system. This process is relatively a time-consuming one, but it wouldn’t delay any real-time call since it only runs few times and barely effects the operation of multi-agent system (the real-time part).In order to verify the model, this article compares the waiting time of emergency calls and water accumulation between the traditional way and bi-layer model in different scenarios. Experimental results show that the bi-layer model can significantly reduce the waiting time of emergency call and decrease the peak water accumulation volume. |
