Emergency department modeling & analysis using Petri nets

Narayanan Varadarajan, Purdue University

Abstract

Emergency department overcrowding has become a major issue over the past decade. Improving efficiency of ED operations has been one of the key objectives of hospitals. This research uses discrete event modeling to model the Emergency Department (ED). Patient flow models for the ED and fast track departments are developed using Petri Nets. Petri Nets are a popular graphical and analytical tool to model discrete event systems. The models enable us to analyze patient flow in the emergency department. A classification of hospital operating states into desirable and divert states is done. Treating the desirable state as an initial state and the divert state as final state; we address the problem of generating the minimum feasible transition firing sequences in Petri nets. Generating legal transition firing sequence from two given Petri net markings is one of the important problems studied in Petri Nets. A two-step mixed-integer programming formulation is used to generate the minimum sequence of events/transitions leading from the desirable state to divert state. In the first step we generate the transition vector that contains the minimum number of transitions that need to fire to move the system from initial state to final state. In the second step we use the initial state, final state and the transition vector generated to obtain a feasible firing sequence. The sequence so obtained identifies the critical events that could lead the hospital to a divert state. The goal of this ongoing research is to develop control policies that enable hospitals to function within certain desirable states and prevent them from reaching undesirable configurations such as an ambulance divert.

Degree

M.S.I.E.

Advisors

Lawley, Purdue University.

Subject Area

Industrial engineering

COinS