Developed by researchers at Ashoka University The tool is also being used to predict Covid-19 in collaboration with global technology company Thoughtworks in Sonipat, Haryana.
It also allows the addition of interventions such as lockdown, vaccine coverage levels to determine the effects of the disease.
The scientists said the model would be made available with open access, which would help governments, NGOs and researchers to intervene and test results through this simulation.
Gautam Mann, professor of physics and biology at Ashoka University, said, “Bharatsim allows us to evaluate the impact of an infectious disease epidemic at the most granular level, because the basic unit is the individual.”
“Imagine a map of India on which many individuals revolve, like a strategy game. These individuals are not all the same: everyone has been carefully created using machine learning techniques to become a real person, a family, a workplace.” Demographic characteristics, “Mann told PTI.
It is developed on the basis of a combination of various large scale surveys like Census, India Human Development Survey (IHDS) etc.
Because each individual is modeled while maintaining the statistical properties of the population, the differential effects of the disease at different ages, the effects of co-morbidities in increasing the risk of adverse outcomes and so much more, the researchers said.
Using the model, researchers can also explore a variety of scenarios, such as the impact of interventions, including lockdowns and school closures.
Debayan Gupta, an assistant professor of computer science at Ashoka University, added, “The model also includes geographical information, so we can see the impact of the disease in different areas like city wards.”
“The model uses high-performance computing infrastructure to drive this large-scale simulation of cities and states with their real-size populations. Mill laptops, “Gupta told PTI.
The output of the system is fed into the “Visualization Engine”, which helps in quick analysis and retrieval of information, otherwise what would be a huge trash of data.
Together, the researchers hope that it will prove to be a powerful aid in understanding the various circumstances associated with extreme granularity.
“For example, the model may allow us to examine the level at which re-infection is ‘critical.’ Or why populations with different age-structures may respond differently to epidemics,” Gupta said.
“It may also allow us to explore the potential impact of more lethal and more transmissible variants. But its most important use is certainly to allow us to compare different strategies for controlling or reducing the epidemic,” he added.
All previous disease models for Covid-19 in India are based on ‘compartmental’ descriptions that form specific assumptions about how people interact with each other to spread the disease.
Mann said these assumptions are not very real.
“Community structure will help in better ways to understand how the rate and pattern of disease spread may change. This is an area in which Bharatsim has a distinct advantage over other methods,” he said.
BharatSim allows individuals to change how they behave, for example, how they may choose to reduce their contact with others, or how some of them may wear masks or violate isolation rules.
Explaining how mathematical modeling works, Mann noted that no model can predict the new COVID-19 wave.
“Multiple scenarios for what a model can do is exploring how a new type that spreads more easily among people but against which previous vaccinations provide little protection, spreads,” he said.
“By comparing the predictions with the data in real time, we can stay ahead of the way the disease spreads, even in the early stages where little is known about the new species,” the scientist explained.