Scheduling Patients in Hospitals Based on Multi-agent Systems

Abstract

Abstract. Scheduling patients in a hospital is a challenging issue as it calls for a sustainable architecture and an effective scheduling scheme that can dynamically schedule the available resources. The objectives of this paper are to propose a viable and systematic approach to develop a distributed cooperative problem solver for scheduling patients based on MAS to minimize the patient stay in a hospital. Our solution approach combines contract net protocol (CNP), multi-agent system architecture, process specification model based on time Petri net (TPN) and optimization theories. Our system first automatically formulate patient scheduling problem based on the TPN models. The patient scheduling problem is decomposed into a number of subproblems that are solved by cooperative agents. A collaborative algorithm is invoked by individual agents to optimize the schedules locally based on a problem formulation automatically obtained by the Petri net models. To realize our solution methodology, we proposed a scheme for publication and discovery of agents' services based on FIPA compliant multi-agent system platform. We develop a scheduling system to solve the dynamic patient scheduling problem.

Literature review on multi-appointment scheduling problems in hospitals

“…To keep the content in line with the scope of this paper, we decided to exclude scheduling problems in research areas other than patient scheduling in hospitals. Book chapter [64,88,98,101] Starting from this initial set of papers, we reviewed both the papers that cited and that were cited by this initial set of papers. Papers (both published and unpublished) obtained by personal communication were also added if they matched the aforementioned criteria.…”

“…Resources are used for treatment purposes [15,22,32,36,56,65,79,82,86,88,101,110,111,129,138] Resources are used for diagnostic purposes [5,31,39,40,41,42,61,66,76,78,85,92,99,100,130,131,137] Some resources are used for diagnostic purposes, while others are used for treatment purposes [6,17,28,45,48,52,64,74,75,102,119,124,134] Resources are not explicitly defined [4,43,44,63,77,…”

“…Minimize access time (or percentage of patients who cannot be seen before a certain deadline) [17,28,32,88,100,101,102,111] Minimize the idle time of resources [15,17,31,74,82,85,89,92] Maximize satisfaction [4,39,44,92,97,98] Minimize the time to complete all tasks (or minimize the waiting time between two consecutive steps) [5,6,17,31,32,40,41,42,43,48,63,64,65,66,75,76,78,85,88,89,99,100,101,102,111,114,124,…”

“…Online scheduling [6,17,28,42,44,45,63,77,78,79,82,89,92,97,98,99,124,130,131] Offline scheduling [5,15,22,31,32,36,39,40,41,48,52,56,61,64,65,66,74,75,76,85,86,88,96,99,100,101,102,105,110,111,114,119,129,…”

“…Tested with fictional data [5,15,22,42,43,63,64,75,77,78,89,92,97,98,114,119,130,138] Tested with real data, but not applied in practice [4,5,6,17,28,31,32,36,39,40,41,44,48,52,56,66,74,79,85,88,99,100,101,102,105,111,124,129,134,137] Tested in practice, ongoing or not permanently implemented [65,82,86] Teste...…”