Overview

This repository is an effort to standardize the interface of the generators in optimization libraries such as Xopt, optimas, libEnsemble, rsopt.

The objective of this effort is for these different libraries to be able to use each other's generators with little effort.

Example: using Xopt generators in optimas

Definitions

• Generator:

  A generator is an object that recommends points to be evaluated in an optimization. It can also receive data (evaluations from past or on-going optimization), which helps it make more informed recommendations.

  Note: The generator does not orchestrate the overall optimization (e.g. dispatch evaluations, etc.). As such, it is distinct from libEnsemble's gen_f function, and is not itself "workflow" software.

  *Examples:

  • Xopt: here is the generator for the Nelder-Mead method. All Xopt generators implement the methods generate (i.e. make recommendations) and add_data (i.e. receive data).
  • optimas: here is the base class for all generators. It implements the methods ask (i.e. make recommendations) and tell (i.e. receive data).

Standardization

Each type of generator (e.g., Nelder-Nead, different flavors of GA, BO, etc.) will be a Python class that defines the following methods:

• Constructor: __init__(self, variables: Dict[str,List[float]], objectives: Dict[str,str], *args, **kwargs):

  The contructor has two mandatory arguments:

  • objectives is a dictionary that lists the objectives to be optimized for. Each objective is a floating point number (objectives are scalars).

    • The keys of this dictionary are the names of the objective. (The same names have to be used in the dictionaries passed to tell.)
    • The values can either be 'MINIMIZE' or 'MAXIMIZE' to indicate whether the objective is to be maximized or minimized.

  • variables is a dictionary that lists the quantities that the generator can vary in order to optimize (i.e. either maximize or minimize) the objectives. Each variable is a floating point number (variables are scalars).

    • The keys of this dictionary are the names of the variables. (The same names have to be used in the dictionaries passed to tell, and are used in the dictionaries returned by tell.)
    • The values are lists of two elements that specify the range of each variable.

  The constructor will also include variable positional and keyword arguments to accommodate the different options that each type of generator has.

  Examples:

  >>> generator = NelderMead( variables={"x": [-5.0, 5.0], "y": [-3.0, 2.0]}, objectives={"f": "MAXIMIZE"})

• ask(num_points: Optional[int] = None) -> List[Dict]:

  Returns set of points in the input space, to be evaluated next. Each element of the list is a separate point. Keys of the dictionary include the name of each input variable specified in the constructor. Values of the dictionaries are scalars.

  In addition, some generators can generate a unique identification number for each point that they generate. In that case, this identification number appears in the dictionary under the key "_id". When a generator produces an identification number, it is important that the identification number is included in the corresponding dictionary passed to this generator in tell (under the same key: "_id").

  • When num_points is not passed: the generator decides how many points to return. Different generators will return different number of points, by default. For instance, the simplex would return 1 or 3 points. A genetic algorithm could return the whole population. Batched Bayesian optimization would return the batch size (i.e., number of points that can be processed in parallel), which would be specified in the constructor.

  • When it is passed: the generator should return exactly this number of points, or raise a error ValueError if it is unable to. If the user is flexible about the number of points, it should simply not pass num_points.

  Examples:

  >>> generator.ask(100)  # too many points
  ValueError

  >>> generator.ask()
  [{"x": 1.2, "y": 0.8}, {"x": -0.2, "y": 0.4}, {"x": 4.3, "y": -0.1}]

• tell(points: List[Dict[str,Any]]):

  Feeds data (past evaluations) to the generator. Each element of the list is a separate point. Keys of the dictionary must include to the name of each variable and objective specified in the contructor.

  Example:

  >>> point = generator.ask(1)
  >>> point
  [{"x": 1, "y": 1}]
  >>> point["f"] = objective(point)
  >>> point
  [{"x": 1, "y": 1, "f": 2}]
  >>> generator.tell(point)