Using LP-Based Heuristics for Solving Large-Scale Integer Problems

Description

Here we will focus on a simple principle on analyzing an MILP problem based on its LP-relaxation. Instead of just applying some rounding heuristic, we will test some key constraints using the variable values from the LP-relaxation and based on the test results fix (or eliminate) these from the consecutive MILP problem instance. We will show that by following the simple principles the MIP solution can be made 2-5 times faster with less than an average of 1% loss of optimality.

Period	17 Nov 2022
Event title	AIChE Annual Meeting
Event type	Conference
Location	Phoenix, United States, ArizonaShow on map
Degree of Recognition	International