D-Wave Optimization

0.6.12 verified Fri May 01 auth: no python

Enables the formulation of nonlinear models for industrial optimization problems using D-Wave quantum and hybrid solvers. Current version 0.6.12, released approximately monthly.

pip install dwave-optimization

Common errors

error ImportError: No module named 'dwave_optimization' ↓

cause Incorrect import path; package uses dot notation.

fix

Use 'from dwave.optimization import Model' (dot, not underscore).

error ModuleNotFoundError: No module named 'dwave.system' ↓

cause Sampler is in dwave-system, which is not installed by default.

fix

Install the missing package: pip install dwave-system

error AttributeError: 'Model' object has no attribute 'sample' ↓

cause The Model class does not have a sample method; use a sampler from dwave.system.

fix

Use DWaveSampler from dwave.system to sample the model.

Warnings

gotcha The Model class is imported from dwave.optimization, not dwave_optimization. Use dot notation. ↓

fix from dwave.optimization import Model

gotcha Symbols are defined and added to the model implicitly upon creation. Do not manually add symbols. ↓

fix x = model.integer(...) — the symbol is automatically registered.

deprecated The use of 'model.minimize()' with a list instead of a sum expression is deprecated. ↓

fix Use sum(x) or model.minimize(reduce(lambda a,b: a+b, symbols, 0))

gotcha Samplers are not included; you must install dwave-system (or dwave-neal for local QUBO sampling). ↓

fix pip install dwave-system

breaking In version 0.6.0, the symbol creation API changed: use model.integer(), model.binary(), model.real() instead of model.symbol(type=...). ↓

fix Replace model.symbol('Integer') with model.integer(...).

Imports

Model

wrong

from dwave_optimization import Model

correct

from dwave.optimization import Model

Package uses dot notation, not underscore.

DWaveSampler

wrong

from dwave.optimization.samplers import DWaveSampler

correct

from dwave.system import DWaveSampler

Sampler is in dwave-system, not in dwave-optimization.

Symbol
```
from dwave.optimization import Symbol
```

Quickstart

Basic nonlinear model creation and sampling via D-Wave hybrid solver.

from dwave.optimization import Model
import os

# Create a model with a symbol
model = Model()
x = model.integer(0, 10, num=3)
model.minimize(sum(x))
# Submit to a solver (requires Leap API token)
token = os.environ.get('DWAVE_API_TOKEN', '')
if token:
    from dwave.system import DWaveSampler
    sampler = DWaveSampler(token=token)
    sampleset = sampler.sample(model)
    print(sampleset.first)
else:
    print('DWAVE_API_TOKEN not set; install dwave-neal for local sampling from dimod')