Module: tf.contrib.constrained_optimization

A library for performing constrained optimization in TensorFlow.

Classes

class AdditiveExternalRegretOptimizer: A ConstrainedOptimizer based on external-regret minimization.

class AdditiveSwapRegretOptimizer: A ConstrainedOptimizer based on swap-regret minimization.

class ConstrainedMinimizationProblem: Abstract class representing a ConstrainedMinimizationProblem.

class ConstrainedOptimizer: Base class representing a constrained optimizer.

class MultiplicativeSwapRegretOptimizer: A ConstrainedOptimizer based on swap-regret minimization.

Functions

find_best_candidate_distribution(...): Finds a distribution minimizing an objective subject to constraints.

find_best_candidate_index(...): Heuristically finds the best candidate solution to a constrained problem.