Design and analysis of a cloud-based epidemic simulation framework

In this paper, we present a cloud-based epidemic simulation framework which is designed for simulating the spread of epidemic diseases globally. We enhance the SEIR, a popular global equation model for infection diseases study. The original SEIR model takes long to generate results. Hence, we modify SEIR by adding the level of severity into each of the epidemic state. Therefore, the time period is used instead of the rate of occurrences in the simulation which dramatically reduces the search space in parameter estimations while maintains the same level of accuracy. Moreover, we enable the modified SEIR model to be parallelizable using a multiple loop decomposition method. It partitions and distributes the workload of the epidemic simulation onto multiple compute resources concurrently. Experiments have been conducted using multiple virtual machines in a cloud environment and the results show that the performance can be improved substantially.