Operations Research groups solve the toughest optimization problems both inside and outside Google.
About the team
Operations Research groups are involved in many areas throughout Google, running the gamut from fundamental research to enterprise-grade engineering. We are software engineers, research scientists, and data scientists who use integer programming, linear programming, constraint programming, and graph algorithms to solve problems at scale.
Across Google, Operations Research tackles challenges in areas as diverse as transportation, search, natural language understanding, machine vision, datacenter design, and robotics. With a strong commitment to open source, we're actively involved in helping solve problems outside Google as well in areas such as aviation and health care.
Team focus summaries
Given a fleet of vehicles and a set of tasks to be performed, what are good routes to take? Such problems crop up frequently in government, logistics, manufacturing, and retail.
Flow and Graph Algorithms
Networks are at the core of many engineering problems. How can you maximize the throughput of a computer network? How can you load-balance scarce resources? Algorithms like max flow and min-cost flow provide the starting points for systems that need to move items through a complex network.
Operations Research began with a seemingly simple question: how can you solve a large set of linear inequalities as efficiently as possible? Research continues to this day.
Many optimization problems involve discrete variables: people, places, and things. A mixed-integer programming problem generalizes linear programming to include discrete variables, with applications to supply chain management, scheduling, bin-packing problems, and much more.
It's often convenient to express optimization problems simply as a set of constraints, variables, and a function to be minimized or maximized. Constraint propagation, aided by heuristics and local search, are used to manage exponentially large search trees.
Advanced analytics permeates work at Google, making the multitechnology giant a "candy store for O.R. practitioners".
Google's open source linear programming solver was used to solve a storied problem faster than ever before, culminating in lunch.