Seeker 1.1 Released

Seeker 1.1 Released

InsideOpt Seeker just got even better. Some highlights: 

  • Improved memory efficiency
  • Faster handling of permutation variables
  • Tenfold better tuning capabilities


As we rolled out Seeker 1.0 half a year ago, many potential clients asked for a comparison with competing solvers on a standard optimization benchmark. For OR scientists, it is clear that any comparison between different algorithm is always a comparison of apples and oranges as the model must fit the solver. That is why we always encourage our clients to try Seeker on their problems, and we help building evaluation prototypes.

However, we understand that you are still interested in such a comparison. So we compared Seeker 1.1 with Gurobi and Hexaly on the Quadratic Assignment Problem (QAP). The QAP is a network design problem that asks for the minimization of transportation costs when distributing n facilities to n given locations. The QAP can be modeled as a standard integer program and is one of the classical OR problems for which we have standard benchmarks. We used the Taillard instances from


Below we compare the gap to the best-known solution for each instance after one hour of computation time.


The above provides a decisive qualitative comparison in favor of Seeker 1.1. But how much faster is InsideOpt Seeker?


In this Google Colab, you can try it yourself:

In this program, Seeker is stopped as soon as it improves over the performance of Gurobi (or Hexaly: simply set 'Gurobi = False') after one hour of computation time on an Intel Core i7-6700k processor (4 cores, 4.0GHz, 8MB cache) and 64GB RAM. We obviously cannot assign the machine type in the Google Colab, but it will usually be less powerful than an i7. Moreover, contrary to the competition, Seeker will only have access to one core. Still, when you try it yourself (simply hit 'Run all' under 'Runtime'), you will typically see that Seeker is 300 to 500 times faster than Gurobi and 25 to 40 times faster than Hexaly.

But: Do you really care about the QAP? Or the Taillard instances? Right. That is why you should try Seeker for the problems that your business needs solved. We help set up these initial prototypes so that you can experience the incredible performance of this new solver first hand.

What are you waiting for? Message to get your free evaluation license and let us know how we can help. Our mission is to help you succeed with your optimization tasks.