Research On HMM-based Emergency State Prediction And Resource Scheduling Algorithm

Emergency resource scheduling is one of the neutron directions of the intelligent scheduling system.The purpose of emergency resource scheduling is to be able to deal with emergency events that may occur in life,such as fires and traffic jams.Emergency events have suddenness,variability and uncertainty,which results in emergency dispatch departments not being able to grasp the current state of emergency events in time.Different types of events have different types of resource requirements,which leads to greater complexity in calculating scheduling schemes for different resources.Therefore,in order to solve the above problems,this thesis proposes solutions to the state prediction of changing emergency events and the realtime resource scheduling algorithm of events,as follows:(1)Aiming at the problem that the existing emergency event sequence regression models ignored the low prediction accuracy caused by event variability,this thesis proposes an event state prediction model based on Hidden Markov Model.First,through the data collection and processing of the event situation,the correlation between the change in event demand and the state change of the event is obtained.Second,the event change probability is obtained through the learning phase of the model.Finally,the prediction algorithm solves the problem of predicting the state change sequence of events.As can be seen in the experiments,the model proposed in this thesis has a high accuracy rate for predicting the variability of event state sequences,and is superior to the existing algorithms in both the prediction effect and the prediction of event changes.(2)Aiming at the problem of insufficient dynamic computing capacity of existing resource scheduling algorithms,this thesis proposes a resource scheduling scheme based on improved genetic algorithm.According to the characteristics that different events have different resource types,an adaptive coding method is designed to update the individual feature matrix of the algorithm synchronously under different resource types and quantity requirements,reducing the impact of redundant features on the algorithm.In addition,this thesis introduces resource selection into the algorithm flow,which improves the transportation efficiency in the scheduling process.Compared with the existing algorithms,it shows the dynamic computing power of the proposed algorithm and the superiority of scheduling optimization.