Scribd
Upload
10 views

CODMAV

Uploaded by

Arya Rushipathak
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
10 views

CODMAV

Uploaded by

Arya Rushipathak
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 7

Hospital Optimisation

System
Ananya Singla : PES2UG21CS066
Anushri Anil : PES2UG21CS081
Arya Rushipathak:PES2UG21CS094
Case Study 5
Domain: Hospital Optimization System Objective:

Problem Statement: 1. To develop a mixed integer programming


model for appointment scheduling of
Scheduling of MRIs in a Hospital Setting: An Optimization
patients in hospital and also researchers
Problem who bring samples.
This study deals with an optimization problem that
appears in the context of booking radiology
appointments at the MRI department. Since MRI 2. How to tackle emergencies?
machines are used by all departments, one of the
busiest sections at the hospital is the radiology 3. How to deal with cancellations of
department and it causes long waiting times for appointments?
patients. Additionally, MRIs are used in hospitals for
research studies approved by their ethics 4. To minimize the waiting time of patients
committees. Patients and researchers who bring as well as researchers.
samples for MRI can reduce their waiting time by
planning for operational restrictions such as
emergencies and cancellations.
Factors Analysed
1. Patient Urgency: The urgency of patients' medical conditions and the need for timely access to
MRI scans to support diagnosis and treatment planning.
2. Resource Availability: The availability of MRI machines, qualified staff, and other necessary
resources to perform scheduled scans.
3. Appointment Prioritization: Prioritizing appointments based on medical urgency, patient needs,
and scheduling policies to ensure that critical cases are addressed promptly.
4. Scheduling Constraints: Operational constraints such as working hours, machine maintenance
schedules, and ethical considerations that may impact the scheduling process.
Methodology
● Our approach is to utilize Graph Neural Networks (GNN) to solve the MRI appointment scheduling
problem.
● We begin by visualizing the entire hospital as a heterogeneous graph consisting mainly of three types of
nodes: patients, MRI machines, and researchers.
● MRI machine attributes include availability and type, while patient attributes include name, ID, severity, and
more.
● Next, we aim to generate node embeddings to represent the attributes of patients, MRI machines, and
researchers, which are then fed into the GNN to streamline scheduling.
● In the GNN model, two main steps are followed: message passing, where nodes gather information from
their neighbors, and aggregation, where this information is processed. The embeddings are then updated,
and weights are assigned based on the criticality of emergencies.
● The GNN suggests optimal appointment scheduling considering factors such as patient needs, machine
availability, and scheduling conflicts.
● Mixed Integer Programming (MIP) takes in the suggestions from the GNN and selects the best schedule. It
considers all rules and constraints, such as avoiding appointment overlaps, maximizing machine efficiency,
and responding promptly to emergencies.
● MIP ensures the integration of GNN-derived insights into a cohesive schedule, resulting in an efficient
allocation of resources and timely patient care.
Methodology
We design the graph to include an "Emergency" node connected to separate MRI machines
designated for emergency cases. Now with the existing GNN model, we prioritize assigning
emergency patients to MRI machines connected to the "Emergency" node. Train the GNN model to
learn patterns and relationships in the graph data, enabling it to make informed decisions during
emergency situations.
Message passing and aggregation in Graph Neural Networks (GNNs) allow nodes to exchange
urgency information, adapt schedules, optimize appointments, and learn patterns. By utilizing this
approach, hospitals can prioritize urgent cases, optimize resource allocation, and minimize waiting
time for patients and researchers in appointment scheduling systems.
Project Flow Diagram
References
[1]
https://www.researchgate.net/publication/343236105_A_Mixed_Integer_
Programming_Model_for_Patient_Appointment_in_an_Infusion_Center

[2]

https://link.springer.com/article/10.1007/s10479-023-05287-5

You might also like