Get the latest tech news

Scaling up linear programming with PDLP


ber 20, 2024 Haihao Lu, Research Scientist, Google Research, and Assistant Professor, MIT, and David Applegate, Principal Scientist, Google Research This post describes the award winning product called PDLP, a new first-order method based solver for large-scale linear programming. Classic linear programming (LP) problems are one of the most foundational problems in computer science and operations research.

With extensive applications across numerous sectors of the global economy, such as manufacturing, networking, and other fields, LP has been the cornerstone of mathematical programming and has significantly influenced the development of today’s sophisticated modeling and algorithmic frameworks for data-driven decision making. Hardware-related challenges: Both methods face difficulties leveraging modern computing architectures, such as GPUs or distributed systems, because the sparse matrix factorization step usually requires highly sequential operations. PDLP has demonstrated its power by solving real-world TSP lower bound LP instances of immense scale, encompassing up to 12 billion non-zero entries in the constraint matrix.

Get the Android app

Or read this on Hacker News

Read more on:

Photo of PDLP

PDLP

Photo of linear programming

linear programming