COVision

COVision integrates agent-based models and classical compartmental models (SIR, SEIR) to simulate SARS-CoV-2 transmission dynamics and evaluate public-health interventions for COVID-19.

Key Features:

• Agent-based models (ABM): Simulates heterogeneous individual interactions to capture complex, long-term epidemic dynamics and variability in COVID-19 spread, with demonstrated advantages over classical compartmental models.
• Classical compartmental models (SIR, SEIR) and R0 computation: Implements SIR and SEIR frameworks and computes basic reproduction number (R0) from these models, including analysis of R0 trends under prolonged lockdown durations.
• Adjustable epidemiological and intervention parameters: Supports simulation across lockdown strength, basic reproduction number (R0), asymptomatic spread, testing rate, contact rate, recovery rate, incubation period, and leakage in lockdown measures.
• Model optimization and evaluation: Models are optimized and evaluated using multiple error metrics.

Scientific Applications:

• Policy evaluation: Simulates and evaluates the impact of public-health interventions such as lockdowns, testing protocols, and social interventions on COVID-19 outcomes.
• Scenario analysis under parameter uncertainty: Generates diverse epidemic scenarios by varying epidemiological and intervention parameters to assess possible trajectories and outcomes.
• State-level intervention assessment in India: Assesses effectiveness of policy interventions across different states in India by analyzing R0 and epidemic dynamics.
• Support for strategic planning and monitoring: Provides quantitative projections to inform strategic planning and monitoring of the COVID-19 epidemic.

Methodology:

Uses agent-based models and classical compartmental models (SIR, SEIR); computes R0 from compartmental models; simulates scenarios by adjusting parameters including lockdown strength, R0, asymptomatic spread, testing rate, contact rate, recovery rate, incubation period, and lockdown leakage; models are optimized and evaluated using multiple error metrics.