Add Egor optimizer wrapper

.github/environment.yml

@@ -21,3 +21,4 @@ dependencies:
 # testing
   - parameterized
   - testflo
+  - egobox >=0.37.6, <0.38.0

doc/index.rst

@@ -14,7 +14,7 @@ The list of supported optimizers is shown on the sidebar to the left.
 Of those, the following are not installed by default:
 
 - SNOPT and NLPQLP are proprietary and must be obtained from their respective authors
-- IPOPT and ParOpt must be installed separately
+- IPOPT, ParOpt and Egor must be installed separately
 
 
 pyOptSparse is a fork of `pyOpt <http://www.pyopt.org/>`_.
@@ -70,3 +70,4 @@ To get started, please see the :ref:`install` and the :ref:`quickstart`.
    optimizers/CONMIN
    optimizers/ALPSO
    optimizers/UNO
+   optimizers/Egor

doc/optimizers/Egor.rst

@@ -0,0 +1,50 @@
+.. _egor:
+
+Egor
+====
+
+Egor is the surrogate-based Efficient Global Optimization (EGO) optimizer from the open-source
+`EGObox <https://github.com/relf/EGObox>`_ library.
+
+Egor uses `bayesian optimization <https://en.wikipedia.org/wiki/Bayesian_optimization>`_ techniques
+well-suited to find the global optimum of an expansive-to-evaluate black-box function.
+Basically, it uses a surrogate model to approximate the objective function and
+an infill criterion (aka acquisition function) to guide the search for the optimum.
+
+The pyOptSparse wrapper is derivative-free and targets single-objective, bounded,
+continuous design spaces. Constraint values are passed to Egor with the pyOptSparse
+constraint convention transformed to :math:`c(x) \le 0`.
+
+Installation
+------------
+
+Egor is made available through the `egobox <https://pypi.org/project/egobox/>`_ Python package.
+
+.. prompt::
+
+   $ pip install egobox
+
+``egobox`` is also available via conda-forge::
+
+   $ conda install -c conda-forge egobox
+
+
+Options
+-------
+
+Please refer to the Egor help for a complete listing of options and their default values.
+
+.. prompt::
+
+   $ python
+   >>> import egobox as egx
+   >>> help(egx.Egor)
+   >>> help(egx.GpConfig)
+
+pyoptSparse expects pickable objects while native Egor structures as GpConfig are not pickable.
+To workaround this constraint, the pyOptSparse Egor wrapper uses dictionaries which are accepted by Egor
+to update the default field values of Egor structures.
+Names and default values of the fields are provided in the descriptions below.
+
+.. optionstable:: pyoptsparse.pyEgor.pyEgor.Egor
+   :filename: Egor_options.yaml

doc/optimizers/Egor_options.yaml

@@ -0,0 +1,98 @@
+gp_config:
+  desc: |
+    GpConfig as a dict used by Egor for surrogate model configuration.
+    Main defaults are:
+
+    - ``regr_spec``: 1 (Constant)
+    - ``corr_spec``: 2 (Squared Exponential)
+    - ``kpls_dim``: no PLS dimensionality reduction (otherwise int)
+    - ``n_clusters``: no clustering (otherwise int)
+cstr_tol:
+  desc: |
+    Constraint tolerances list passed to Egor (size n_cstr plus n_fcstr)
+    Default is ``1e-4`` for all constraints.
+n_start:
+  desc: Number of infill optimization runs (best run selected)
+n_doe:
+  desc: Number of initial DOE samples (0 lets Egor choose automatically)
+doe:
+  desc: |
+    Initial DOE array, either x-only or concatenated x and y
+    to be passed as list of lists of floats. If not provided, Egor will generate a DOE automatically.
+infill_strategy:
+  desc: |
+  Infill criterion:
+
+    - 1 = ``EI``,
+    - 2 = ``WB2``,
+    - 3 = ``WB2S``,
+    - 4 = ``LOG_EI`` (default)
+cstr_infill:
+  desc: Enable constrained infill criterion (aka CEI)
+cstr_strategy:
+  desc: |
+    Constraint strategy enum for surrogate constraint handling:
+
+    - 1 = ``MeanConstraint`` (default)
+    - 2 = ``UpperConfidenceBound``
+qei_config:
+  desc: |
+    QEiConfig for batch (qEI) point selection passed as a dict with main keys being:
+
+    - ``batch``: size of batch (int)
+    - ``strategy``: qEI strategy enum
+infill_optimizer:
+  desc: |
+    Internal infill optimizer:
+
+    - 1 = ``COBYLA`` (default)
+    - 2 = ``SLSQP``
+trego:
+  desc: |
+    Enable TREGO (aka Trust Region EGO) algorithm configured with main parameters:
+
+    - ``n_gl_steps``: (nb of global search steps default 1, nb of local search steps default 4)
+    - ``beta``: trust region factor (default 0.9)
+
+coego_n_coop:
+  desc: Number of cooperative groups for CoEGO algorithm
+target:
+  desc: Known objective target used as stopping criterion
+outdir:
+  desc: Output directory for Egor output files (configuration, does, history and warm start search)
+warm_start:
+  desc: Load initial DOE from outdir when enabled
+hot_start:
+  desc: |
+    Egor checkpoint restart parameter to be used in case of fallible environment
+    to continue with the same Egor parameterization till max iterations or timeout is reached.
+failsafe_strategy:
+  desc: |
+    Failure handling enum
+
+    - 1 = ``REJECTION`` (default),
+    - 2 = ``IMPUTATION``,
+    - 3 = ``PROBA OF VIABILITY``
+seed:
+  desc: Seed for random number generator (default -1 for random seed)
+verbose:
+  desc: |
+    Verbosity level for Egor logging
+
+    - 0 = ``error`` (default)
+    - 1 = ``warning``
+    - 2 = ``info``
+    - 3 = ``debug``
+max_iters:
+  desc: Egor iteration budget
+run_info:
+  desc: Optional RunInfo used to pass additional information to Egor (e.g., for logging)
+timeout:
+  desc: Optional timeout in seconds used as sttopping criterion for Egor minimize
+fcstrs:
+  desc: |
+    Optional list of native Egobox function constraints passed directly as fcstrs
+    (instead of pyOptSparse constraints which are metamodelized)
+fcstr_specs:
+  desc: |
+    Optional list of egobox.CstrSpec for function constraints passed as fcstrs

pyoptsparse/__init__.py

@@ -21,6 +21,7 @@
 from .pyNSGA2.pyNSGA2 import NSGA2
 from .pyALPSO.pyALPSO import ALPSO
 from .pyParOpt.ParOpt import ParOpt
+from .pyEgor.pyEgor import Egor
 
 __all__ = [
     "History",
@@ -43,6 +44,7 @@
     "NSGA2",
     "ALPSO",
     "ParOpt",
+    "Egor",
     "testing",
     "list_optimizers",
 ]