PathWise Insights

How High Performance Computing is Addressing Complex Pandemic Challenges

Peter Phillips

President and CEO of PathWise Solutions Group

Pouyan Djahani

Managing Director, PathWise Solutions Group

The push to innovate starts right at the top of our company. The PathWise Solutions Group has a long track record of using Graphics Processing Units (GPUs) and cloud computing to help life insurance companies understand and manage the complex financial guarantee risk embedded in their life insurance retirement products. We are no strangers to high performance computing, stochastic-differential equations, simulations and working with big data on GPUs.

The starting point for the work that follows was a paper published by the Faculty of Medicine at the Imperial College in London in March of 2020. It discussed modeling of the pandemic along with a spotlight on the impact of various policy interventions might have on the course of the actual pandemic.  It generated a lot of publicity and media attention. However, it was not clear to us what tools were used to produce the results, and if they could be replicated. Nevertheless, the references and the presentation itself told a story about risk – one of our key areas of expertise. Then an article appeared in the New York Times, which talked about what the pandemic risk might look like for given vaccine efficacy rates, but not using the rate quoted by Moderna or Pfizer as a basis. This led us to consider looking at the range of possible risk trajectories by modeling through Monte Carlo simulation, to better understand the distribution of possible outcomes suggested by the model dynamics versus just looking at expected outcomes. We also wanted to be able to assess what might happen under different vaccine efficacy rates, and to examine the interplay between different policy decisions over a range of possible outcomes.

What follows is a modeling exercise, and is short examination of pandemic modeling using our high performance computing platform, PathWise, in an area we had never previously explored. I am proud of the team’s efforts here, and I hope it encourages others to examine this interesting field more closely, and to think about using data and analytics to shine a light on important and complex problems faced by society today.

Download the article