Pyoptsparse linear constraint bug.

lincon_bug.py

"""
Equivalent OpenMDAO example.

import numpy as np
import openmdao.api as om

prob = om.Problem()
model = prob.model

model.add_subsystem('comp', om.ExecComp(['y=2.0*x + 100', 'y2=3.0*x', 'obj=sum(-3.0*x**2)']))
model.add_design_var('comp.x', lower=0.01, upper=4.0)
model.add_objective('comp.obj')
model.add_constraint('comp.y', upper=2.0, linear=True)
model.add_constraint('comp.y2', upper=7.0)

prob.driver = om.pyOptSparseDriver(optimizer='SNOPT')

prob.setup()

prob.set_val('comp.x', 0)
prob.run_driver()
"""
import numpy as np

from pyoptsparse import SNOPT, Optimization

LINEAR = True


def objfunc(xdict):
    """Evaluates these equations:
    obj = -3.0 * x**2
    con1 = 2.0 * x + 100
    con2 = 3.0 * x
    """
    x = xdict["x"]
    funcs = {}

    funcs["obj"] = -3.0 * x**2
    funcs["con2"] = 3.0 * x

    if not LINEAR:
        funcs["con1"] = 2.0 * x + 100

    fail = False
    return funcs, fail


def sens(xdict, funcs):
    """Derivatives"""
    x = xdict["x"]
    funcsSens = {}

    funcsSens['obj'] = {"x": -6.0 * x}
    funcsSens['con2'] = {"x": 3.0}

    if not LINEAR:
        funcsSens['con1'] = {"x": 2.0}

    fail = False
    return funcsSens, fail


# Optimization Object
optProb = Optimization("Test1", objfunc)

# Design Variables
optProb.addVarGroup("x", 1, varType="c", lower=0.1, upper=4.0, value=0.0)

# Objective
optProb.addObj("obj")

# Constraint
optProb.addConGroup("con2", 1, upper=7.0, wrt=["x"], linear=False)

# Linear Constraint
if LINEAR:
    jac = {"x": 2.0}
    optProb.addConGroup("con1", 1, upper=2.0, wrt=["x"], linear=LINEAR, jac=jac)
else:
    optProb.addConGroup("con1", 1, upper=2.0, linear=False)

opt = SNOPT()
sol = opt(optProb, sens=sens)
print(sol)