We help our clients wrangle problems that seem computationally intractable; be they due to tremendous amounts of data, or massive search spaces. Our expert algorithm designers solve these problems by drawing upon a range of techniques, from branch-and-bound and dynamic programming to massively parallel programming, mixed integer programming and AI.
Using these optimization techniques coupled with advanced modeling techniques, we have solved complex problems and developed automated optimization algorithms for industrial, business, engineering and public sector operations.
We have also applied the same large-scale optimization techniques to solve problems that would not seem like “operations” research at first. Problems with a large number of individual decisions (called “degrees of freedom” in optimization lingo) occur not only in enterprise operations, but also in engineering design, game playing algorithms, real-time control algorithms, machine operation sequencing and automation, and many other situations.
EXAMPLE PROJECTS
Logistics Optimization
Logistics problems like vehicle routing, container stacking, goods packing, transport scheduling and routing are extraordinarily difficult to optimize. Even though these problems are “NP-complete” — computer science code for very very hard — we can use advanced algorithms and heuristic techniques to get certifiably close to the optimal solution.
Learn more
Parallel Scheduling
Running an algorithm on a large supercomputer is not easy. The algorithm has to be broken up into smaller tasks, and each of these tasks has to be run on individual nodes of the supercomputer. Which task should be run where, when? This is the complex question answered by scheduling algorithms.
Learn more about our scheduling expertise
Public Transport Optimization
What would be the optimal schedule for a particular route? What new routes should be started? Which routes should be modified? Optimizing a fleet of vehicles to best serve the public interest is not easy.
Learn how algorithms help
Pricing
To price products well, various price decisions have to be run into the future, to understand their overall impact. Furthermore, price decisions may be re-taken at future moments, and competitors may change their decisions. The state of the economy may change, impacting demand. Taking the optimal decision now considering all possible future scenarios is a difficult task: and an algorithmic technique called “dynamic programming” helps.
Learn more about pricing and econometrics
Oil & Gas Production Optimization
We designed and built a platform for data-informed and physics-informed simulations of large-scale multi-year oil & gas production projects. The platform integrates diverse data and simulations, runs multiple scenarios, and makes future projections of production decisions and infrastructure development.
Learn more