ecole
ecole copied to clipboard
ds4dm
Extensible Combinatorial Optimization Learning Environments
## Describe the problem or improvement suggested Build wheels for fast and easy installation using pip. ## Describe the solution you would like - [ ] [cibuildwheel](https://github.com/pypa/cibuildwheel) can be used...
## Describe the problem or improvement suggested For researchers to work with Ecole in C++ and Python they have two choices: - 1. Work in a fork of Ecole, which...
### Discussed in https://github.com/ds4dm/ecole/discussions/193 A partial fix has been merged into PySCIPOpt [here](https://github.com/scipopt/PySCIPOpt/pull/505). Still, I understand that one must keep a reference to the PySCIPOpt object that was used to...
## Describe the problem or improvement suggested Right now, the complete history of bounds is stored in the event handler and the integral is recomputed from start before taking a...
## Describe the bug With the current reward function implementation we are unable to declare multiple reward function in a single environment. ## Setting - OS: Any - Compiler: Any...
## Describe the bug A segfault happens randomly when using the Configuring environment as in the tutorial. ## Setting - OS: Linux 5.9.11-200.fc33.x86_64 #1 SMP Tue Nov 24 18:18:01 UTC...
## Describe the problem or improvement suggested Get Ecole working on Windows 🚀 💻 ## Things that need to be done: - [ ] Ship SCIP & pyscipopt on Windows...
For example, the `NodeBipartite` extraction function can only be called when the SCIP stage is `SCIP_STAGE_SOLVING` or later. Following an internal discussion, we have converged on two proposals that would...
## Describe the problem or improvement suggested An instance generator that download a remote archive, extract it, and iterate over the files. ## Describe the solution you would like -...
## Describe the problem or improvement suggested Many simple reward functions have the following pattern ```cpp class SomeReward { void before_reset(...) { metric = 0; } double extract(...) { auto...