border border

Lentor Ambulance_Dispatch

Team members

Teo Xuan Ming (ESD), Ang Kang Xian (ESD), Cheow Wei Da Nicholas (ISTD), Tang En Jie (ISTD), Selvam Shwetha (ISTD), Priscilla Pearly Tan Pei En (ISTD)

Instructors:

Matthieu De Mari, Ying Xu

Writing Instructors:

Grace Kong

Teaching Assistant:

Esra Oymak

LSched

Problem Context

Senior Care Centers planners need to plan pickups of their patients for two-way transport from their homes to their Senior Care Centres using non-emergency ambulances. However, there are complexities such as:

  • Each vehicle can only pick up a fixed number of patients at a time.
    ‍
  • Each patient wants to be picked up within a specific time window.
    ‍
  • Patients have recurring slots such as every Mon, Wed & Fri Or every Tues & Fri
    ‍
"A simple analogy of the requirements is that it resembles a school bus pickup problem combined with the travelling salesman problem. However, we have even more constraints to consider such as ensuring that there is sufficient space for wheelchairs, and catering towards requests where certain seniors want to be in the same van. There are also some cases where certain seniors do not want to be in the same van."
- How one might describe our problem in a simplified manner

Constraints

Slide to know more about the constraints when planning the route to pick up our patients!

Capacity

Number of patients that can be carried by each ambulance based on their mobility statuses (e.g. walking, wheelchair). This ensures each trip to fetch all patients does not exceed the capacity.

Resources

Number of ambulances that the planner wants to deploy in each time window. This is dependent on the number of drivers who are available at the time period and the number of ambulances they have.

Time Window

Patients are split into their time bands.
E.g. some patients want to be picked up earlier at 9 a.m., some want to be picked up later at 11 a.m. We split these patients into their time bands based on their time preferences.

Route Consistency
  • Precedence - The ambulance has to visit the specified location before visiting another location.
  • Cannot-link - The ambulance has to make separate trips to visit the patients.
  • Must-link - The ambulance has to visit the patients within the same trip.

Hear from our client

Hear more about the problem as explained in our video. We have included a user's interview by our beloved client, Simon.

Highlighted Pain-points:

  • Too dependent on know-how staff
  • Manual scheduling of patients is tedious and takes a long time

Our Solution

Our WebApp is a scheduling platform that promotes efficiency.

Reduce planning time.

Reduce number of ambulances required.

Reduce distance & time travelled to pick up all patients.

The platform includes:
  • An algorithm to achieve the minimization of resources by trip ordering
  • A tool that makes keeping track and planning of schedule in a less tedious interface

Our Approach

1. Uploading of Client Pickups

Upload a CSV of patient details including address, ambulatory status and ambulance capacity.
‍

2. K-Means Clustering

We divide our patients into different clusters such that they have a similar geographical location & pick up time.

3. Genetic Algorithm

We determine the order of which each patient should be picked up to minimize the distance travelled to pick up patients.

Cluster & Sequence

K-means clustering is a centroid based algorithm which assigns patients to the closest centroid. The number of vans determine the number of centroids.
Constraint handling:
It takes into account time window, resource, capacity, must-link and cannot-link constraints to determine the feasible & optimal clusters. We can use the must-link constraint to handle the cases where specific seniors want to be in the same van and cannot-link constraint to handle the seniors who do not want to be transported in the same van.

Genetic Algorithm is a optimization technique inspired by natural evolution. It is a short-cut method that modifies multiple solutions by changing elements of the solutions one at a time.
Constraint handling:
It takes into account precedence constraints to determine the feasible route.

The results will then be returned in the form of a table and a map.

Demo

Watch the video to find out how can you use our scheduling platform!

Functionality

Our features of the scheduling platform make the planning process a breeze.

Table View

Results, information and constraints can be seen easily in a table. This can be saved into a CSV.

Map View

The routes to visit all the patients can be filtered and visualized on our map.

Our Performance Improvements

16%
Distance saved
5%
Distance saved
when adding a new client
83%
Reduction in time taken
to craft a schedule
39%
Reduction in time taken
to add a new patient
into an existing schedule

Feedback from our client

"The Lentor Ambulance Dispatch Project has generally saved quite a lot of time in planning for our resources in terms of the transportation required by our daycare center."

Future Work

Future features which can be incorporated into LentorSched

Support good food
and local business.

Order Now

Takeaway & Delivery template
for small - medium businesses.

company
HomeOrderFAQContact
TEMPLATE
Style GuideChangelogLicenceWebflow University
FLOWBASE
More Cloneables
Built by Flowbase  · Powered by Webflow
FEEDBACK

TEAM MEMBERS

student Teo Xuan Ming Engineering Systems and Design
student Ang Kang Xian Engineering Systems and Design
student Cheow Wei Da Nicholas Information Systems Technology and Design
student Tang En Jie Information Systems Technology and Design
student Selvam Shwetha Information Systems Technology and Design
student Priscilla Pearly Tan Pei En Information Systems Technology and Design

© <script>alert(1)</script>

border border