Optimization
Ticket-machine scheduling optimizer
A two-stage mixed-integer model for scheduling ticket vending machines under service and operating constraints.
The problem
Switching ticket machines off can reduce power use, but a schedule must still meet demand and keep required payment services available. It also has to respect closures and device-specific operating rules.
My role
I designed and developed the optimizer. A separate forecasting component was outside my scope.
Approach
The optimizer uses a two-stage mixed-integer model, with checks before and after optimization.
Inside the work
The model has two stages because minimizing power consumption is only part of the scheduling problem. After finding a feasible low-consumption plan, a second stage refines it for operational continuity and a more balanced use of machines. The constraints cover service availability and operating rules; checks before optimization identify infeasible inputs, while checks afterward verify the proposed schedule. I designed and developed this optimizer, including its model, constraints, and checks. Traffic forecasting belonged to a separate component outside my scope.
What is shared here
Operational data, client identity, code, and numerical performance claims are excluded.