Seeker is a powerful primal solver designed to find optimal and near-optimal solutions fast, making it ideal for real-world modeling and decision-making under time constraints. By prioritizing the best possible solutions over traditional bounds, Seeker excels where legacy solvers fall short.

Key Advantages

• Time-Constrained Optimization: Automatically tunes its search strategy for maximum performance based on specific time limits.
• Unprecedented Parallelization: Fully distributed, allowing for massive performance gains when utilizing dozens or even hundreds of workers on a single problem instance.
• Complex Modeling: Direct optimization of non-linear, non-convex, and non-differentiable relationships. Prototype extremely fast by deriving KPIs directly, like in a simulation, eliminating the need for linearization or binarization.
• Multi-Objective Optimization: Optimize multiple KPIs simultaneously without aggregation, strict hierarchies, or constant trade-off rates. This provides practical solutions, even for infeasible problems.
• Decision-Making Under Uncertainty: Built for stochastic optimization. Optimize for expected performance, quantiles, Conditional Value at Risk (CVaR), or risk probability, based on the award-winning work of its creators.

Proven Results

See concrete examples, including code that runs on Google Colab, on the InsideOpt Webpages. Highlights include:

• A quadratic assignment example where Seeker demonstrated a speedup of over 30x-1,000x compared to others.
• In a pricing and distribution case, Seeker delivered $186 million more for the clients.