After Seeker 1.0, 1.1, and 1.2, we now make our fourth release of Seeker!
What is new:
1. Dynamic Variable Priorities. You can now hint the solver to look primarily at variables that, under the current assignment, appear most important to achieve improvements. The priority values are dynamic. You define how to compute the priorities dynamically!
2. Dynamic Variable Blocks. You can tell the solver, via a dynamic Boolean term, not to look at some variables under the current assignment but to focus on the unblocked variables only.
3. Nested Equation Systems. Just as you can infer the value of continuous variables by using nested linear programs, you can now determine their values by solving nested equation systems. No need to search for values that can be computed efficiently, right?
4. Delayed Evaluation. For expensive terms, such as nested LPs and EQs, you can tell the solver under what conditions a recomputation is warranted and when it may skip it, using a dynamic Boolean term.
How to get started:
pip install insideopt-seeker
Or use our free version on Google Colab:
pip install seekerdemo
Example: https://lnkd.in/ev7TS3NA
Documentation: https://lnkd.in/ezbK-xFv
YouTube Tutorials: https://lnkd.in/e9EXzR6Z
Happy Optimizing!
