A hybrid simulation model for preparedness

Abstract

Defence allocates considerable resources to maintain preparedness. Preparedness is the capacity of defence to sustainably provide forces that are able to accomplish government-directed tasks. Defence Preparedness Requirements plans are produced every year and agreed upon by the chiefs of services. Each service commits to deliver Defence Elements to perform specific military roles in certain numbers and at certain readiness notices. In light of this, the question arises: how well do these preparedness plans perform in reality? That is to say, how well prepared are we to perform disaster relief operations, military conflicts, and other operations, within reasonable warning times? In order to address this question, a non-linear dynamic probabilistic model of preparedness called DyPSim, has been developed and implemented in Java. DyPSim is a hybrid simulation model that quantitatively determines Defence Element availability for a given force structure subject to the demands of a series of military operations. We have employed the concept of supply and demand to model the preparation of Defence Elements for deployment, and their subsequent deployment on military operations. DyPSim combines continuous processes described by differential equations (the supply model) with stochastic discrete events (the demand model). We present the results of running an unclassified dataset through the model for verification purposes. The system performance metric has been defined as the ratio of successful events divided by failed events to the total number of events during the simulation period. Events are judged to be successful or failed on the basis of their level of resourcing and the timeliness of the resourcing. System success as a percentage was measured by the number of successful events during 30 years of simulation time compared to the total number of events over this period and averaged out over 30 simulation runs. The system failure rate was measured in an equivalent manner. A tipping point can be seen in Figure A1 above which the number of failed events will exceed the number of successful events. Lack of resources in the appropriate time frames due to the spread of resources across different events will cause more failures than successes. We now have two ways to assess the performance of a force structure: we can monitor the success or failure of key events, and we can monitor the system failure to success ratio. Figure A2 shows the success and failure trend as a function of the number of concurrent events. The obtained results demonstrate force structure suitability as a function of the concurrent event profile in the context of preparedness requirements.