Comparing Decentralized Algorithms for Dynamic Task Sharing among Agents with Limited Resources

Dynamic task sharing among organizations is attracting attention in research as the COVID-19 pandemic continues and many hospitals are filled with patients. Here, the task is, for example, the treatment of patients requiring resources such as beds and medical staffs for certain periods of time. This task must be started before its completion becomes overdue. However, because these tasks are created dynamically, the task-handling cannot be scheduled in advance; in addition, these tasks are handled by multiple independent organizations. Because the distribution of tasks are not uniform across the organizations, some tasks should be transferred to unoccupied organizations. In this study, we present and compare some decentralized algorithms for dynamic task sharing among organizations with limited resources. We aim to minimize the number of unstarted tasks within the time limit.