AVL-based & Station-based Dispatch: What the Differences Are and How to Simulate Complex Dispatch Policies
How much time is AVL saving you?
Does AVL provide a good return on investment?
What exactly does AVL do to help?
Deccan International, the industry leader in Fire & EMS deployment analysis for more than 20 years, can provide answers to these questions and more with ARI Simulator, a novel simulation tool.
That’s Apparatus Response and Incident Simulator.
Before getting into how ARI Simulator can help, let’s first examine what AVL-based dispatch is and how it differs from what many Fire & EMS departments currently use: station-based dispatch.
Figure 1 illustrates the difference between station-based and AVL-based dispatch. As a result of AVL-based dispatch, we expect the response times to decrease as units closer to the incident are dispatched. This is exactly what we see in our model, as shown in Figure 2. However, there is surprisingly only a very small difference (2-3 seconds for both the average and 90th percentile) between the two dispatch policies. On further investigation, we hypothesized that this could be because of the low number of incidents for this example client.
To test our hypothesis, we doubled the number of incidents. As expected, we saw a greater difference between the scores of station-based and AVL-based dispatch. As shown in Figure 3, AVL saves 8 seconds on average and 4 seconds on the 90th percentile. These results motivated us to further increase the number of incidents.
In Figure 4, we see that the difference between station-based and AVL-based dispatch grows with increasing number of incidents, to the point where it is about 35 seconds on average and 30 seconds for the 90th percentile for close to 55,500 incidents or seven times the actual number of incidents. AVL-based dispatch, therefore, provides significant savings in time when there are a high number of incidents. Does this mean that these results are applicable for any region with around 55500 incidents? No, as the time saved by AVL is also a function of unit availability and the spatial distribution of incidents, among other factors.
In Figure 5, we see that the average unit availability corresponding to 8,039 incidents is 96.48%, but is only 75.39% for 55,563 incidents. This gives our clients a better picture when trying to determine how much time AVL-based dispatch can save them – if the average unit availability is around 75% and there are close to 55,000 incidents per year, a difference of about 30 seconds can be expected in the 90th percentile response times. However, as indicated previously, this is not the complete picture as the difference is a function of many other factors which differ from region to region. Further explanation as to why we see a greater difference with increasing number of incidents is shown by the curve representing the percentage of responses using AVL dispatch. While only 1% of the responses used AVL-based dispatch with 8039 incidents, more than 10% of the responses used AVL-based dispatch with 55,563 incidents.
We have established that AVL-based dispatch results in significantly better response times in a region with a high number of incidents. But, why is AVL-based dispatch better? To find the answer to this question, we segmented our response times into travel and turnout times. As we are dispatching units that are closer to the incident, we expect the travel times to account for a significant portion of the difference we see in the average response times between station-based and AVL-based dispatch. Surprisingly, we notice that most of the saving comes from lower turnout times rather than travel times, as shown in Figure 6. This is largely because a unit takes longer to physically move after it has received a notification for dispatch when it is at a station compared to when it is returning to a station.
AVL-based dispatch is only one of the complex dispatch policies that can be modelled by ARI Simulator. So, what is ARI Simulator?
ARI Simulator is a simulation model that replicates the complex processes followed by Fire & EMS departments. It can be thought of as a digital lab where you can conduct different experiments to determine how changes can impact your system.
ARI Simulator can be used by Fire and EMS departments, as well as communication centers or CAD vendors all over the world. Fire & EMS departments, for example, can use the ARI Simulator to estimate the impact of complex dispatch policies before implementing changes in the field, while CAD vendors or their customers can use it to determine the sensitivity of the system to new response plans before going live.
ARI Simulator is currently available on a consulting basis, while the stand-alone product is expected to be available by early 2017.
Watch this space for more analysis and discussion on how ARI Simulator and other Deccan International solutions can help the public safety industry.
In order to carry out proper analyses, we must first ensure that our model accurately represents reality. Figure A.1 shows the average, 90th percentile and the 48-second percentage compliant scores when looking at first emergent engine response times for an example client. The “CAD Analyst” column displays scores from our application CAD Analyst, our “what is” historical performance application, while the “ARI Simulator” column displays scores from our simulation model. It is evident that the scores are similar, especially the 90th percentile response times – an important metric in the public safety industry. This gives us confidence that our model represents the system well and allows us to perform analyses with confidence.