22. Microsimulation
In this section we are going to describe how to run simulations of the Negligence-Defendant-Microsim model using Microsimulation. Microsimulation runs sets of trials through the same model. We could run 1000 different trials of the Negligence-Defendant-Microsim.trex model and the outputs allow you to assess the overall decision of the model based on the aggregated set of outcomes.
Microsimulation uses Monte Carlo Simulation. Some other names this analysis is also referred to as:
-
Microsimulation - a common name in this text
-
1st-order trials
-
Random walk
-
Monte Carlo Simulation (we try to avoid using this term in this context because Monte Carlo Simulation is a broader term referring to several types of simulation)
When you run Microsimulation, you can set up distributions which provide variance around the base case input value. When you run trials through the model, each trial will take a sample from the distributions and run the model using that sample value. In the case of the Negligence model, this will give us the ability to also look at each individual simulation.
There are some other things which are also important to consider when running Microsimulation. This includes:
-
Learn about sensitivity analysis on trial level simulation models: Probabilistic Sensitivity Analysis and Microsimulation
-
Run multiple simulations in a batch to determine if your results are stable: Robustness of Simulation Results