Optimizers (qiskit.aqua.components.optimizers)

Aqua contains a variety of classical optimizers for use by quantum variational algorithms, such as VQE. Logically, these optimizers can be divided into two categories:

Local Optimizers

Given an optimization problem, a local optimizer is a function that attempts to find an optimal value within the neighboring set of a candidate solution.

Global Optimizers

Given an optimization problem, a global optimizer is a function that attempts to find an optimal value among all possible solutions.

Optimizer Base Class

Optimizer

Base class for optimization algorithm.

Local Optimizers

ADAM	Adam and AMSGRAD optimizer.
AQGD	Analytic Quantum Gradient Descent (AQGD) optimizer.
CG	Conjugate Gradient optimizer.
COBYLA	Constrained Optimization By Linear Approximation optimizer.
L_BFGS_B	Limited-memory BFGS Bound optimizer.
GSLS	Gaussian-smoothed Line Search.
NELDER_MEAD	Nelder-Mead optimizer.
NFT	Nakanishi-Fujii-Todo algorithm.
P_BFGS	Parallelized Limited-memory BFGS optimizer.
POWELL	Powell optimizer.
SLSQP	Sequential Least SQuares Programming optimizer.
SPSA	Simultaneous Perturbation Stochastic Approximation (SPSA) optimizer.
TNC	Truncated Newton (TNC) optimizer.

Global Optimizers

The global optimizers here all use NLopt for their core function and can only be used if their dependent NLopt package is manually installed. See the following section for installation instructions.

• Installing NLopt

The global optimizers are as follows:

CRS	Controlled Random Search (CRS) with local mutation optimizer.
DIRECT_L	DIviding RECTangles Locally-biased optimizer.
DIRECT_L_RAND	DIviding RECTangles Locally-biased Randomized optimizer.
ESCH	ESCH evolutionary optimizer.
ISRES	Improved Stochastic Ranking Evolution Strategy optimizer.