Changeset 748
- Timestamp:
- 09/20/2010 07:45:24 PM (2 years ago)
- Files:
-
- PythonPackages/OpenOpt/openopt/doc/badlyScaled.py (modified) (1 diff)
- PythonPackages/OpenOpt/openopt/examples/nlp_1.py (modified) (1 diff)
- PythonPackages/OpenOpt/openopt/examples/nlp_11.py (modified) (1 diff)
- PythonPackages/OpenOpt/openopt/examples/nlp_3.py (modified) (1 diff)
- PythonPackages/OpenOpt/openopt/examples/nlp_ALGENCAN.py (modified) (1 diff)
- PythonPackages/OpenOpt/openopt/examples/nlp_bench_2.py (modified) (2 diffs)
- PythonPackages/OpenOpt/openopt/tests/nlpLC.py (modified) (1 diff)
Legend:
- Unmodified
- Added
- Removed
- Modified
- Copied
- Moved
PythonPackages/OpenOpt/openopt/doc/badlyScaled.py
r170 r748 10 10 x0 = [-4,4] 11 11 # even modification of stop criteria can't help to achieve the desired solution: 12 someModifiedStopCriteria = {'g radtol': 1e-15, 'ftol': 1e-13, 'xtol': 1e-13, 'maxIter': 1e3}12 someModifiedStopCriteria = {'gtol': 1e-15, 'ftol': 1e-13, 'xtol': 1e-13, 'maxIter': 1e3} 13 13 14 14 # using default diffInt = 1e-7 is inappropriate: PythonPackages/OpenOpt/openopt/examples/nlp_1.py
r208 r748 84 84 85 85 # If you use solver algencan, NB! - it ignores xtol and ftol; using maxTime, maxCPUTime, maxIter, maxFunEvals, fEnough is recommended. 86 # Note that in algencan g radtol means norm of projected gradient of the Augmented Lagrangian86 # Note that in algencan gtol means norm of projected gradient of the Augmented Lagrangian 87 87 # so it should be something like 1e-3...1e-5 88 88 gtol = 1e-7 # (default gtol = 1e-6) PythonPackages/OpenOpt/openopt/examples/nlp_11.py
r250 r748 95 95 96 96 # If you use solver algencan, NB! - it ignores xtol and ftol; using maxTime, maxCPUTime, maxIter, maxFunEvals, fEnough is recommended. 97 # Note that in algencan g radtol means norm of projected gradient of the Augmented Lagrangian97 # Note that in algencan gtol means norm of projected gradient of the Augmented Lagrangian 98 98 # so it should be something like 1e-3...1e-5 99 99 gtol = 1e-7 # (default gtol = 1e-6) PythonPackages/OpenOpt/openopt/examples/nlp_3.py
r170 r748 64 64 p = NLP(f, x0, c=c, h=h, lb = lb, ub = ub, ftol = 1e-6, maxFunEvals = 1e7, maxIter = 1220, plot = 1, color = color, iprint = 0, legend = [solvers[j]], show= False, xlabel='time', goal='maximum', name='nlp3') 65 65 if solver == 'algencan': 66 p.g radtol = 1e-166 p.gtol = 1e-1 67 67 elif solver == 'ralg': 68 68 p.debug = 1 PythonPackages/OpenOpt/openopt/examples/nlp_ALGENCAN.py
r249 r748 87 87 p.contol = 1e-3 # required constraints tolerance, default for NLP is 1e-6 88 88 89 # for ALGENCAN solver g radtol is the only one stop criterium connected to openopt89 # for ALGENCAN solver gtol is the only one stop criterium connected to openopt 90 90 # (except maxfun, maxiter) 91 # Note that in ALGENCAN g radtol means norm of projected gradient of the Augmented Lagrangian91 # Note that in ALGENCAN gtol means norm of projected gradient of the Augmented Lagrangian 92 92 # so it should be something like 1e-3...1e-5 93 p.g radtol = 1e-5 # gradient stop criterium (default for NLP is 1e-6)93 p.gtol = 1e-5 # gradient stop criterium (default for NLP is 1e-6) 94 94 95 95 PythonPackages/OpenOpt/openopt/examples/nlp_bench_2.py
r170 r748 38 38 lb[3] = 5.5 39 39 ub[4] = 4.5 40 g radtol=1e-140 gtol=1e-1 41 41 ftol = 1e-6 42 42 xtol = 1e-6 … … 60 60 solver = solvers[j] 61 61 color = colors[j] 62 p = NLP(f, x0, name = 'bench2', df = df, c=c, dc = dc, h=h, dh = dh, lb = lb, ub = ub, g radtol=gradtol, ftol = ftol, maxFunEvals = 1e7, maxIter = maxIter, maxTime = maxTime, plot = 1, color = color, iprint = 10, legend = [solvers[j]], show=False, contol = contol)63 # p = NLP(f, x0, name = 'bench2', df = df, c=c, dc = dc, lb = lb, ub = ub, g radtol=gradtol, ftol = ftol, maxFunEvals = 1e7, maxIter = 1e4, maxTime = maxTime, plot = 1, color = color, iprint = 0, legend = [solvers[j]], show=False, contol = contol)62 p = NLP(f, x0, name = 'bench2', df = df, c=c, dc = dc, h=h, dh = dh, lb = lb, ub = ub, gtol=gtol, ftol = ftol, maxFunEvals = 1e7, maxIter = maxIter, maxTime = maxTime, plot = 1, color = color, iprint = 10, legend = [solvers[j]], show=False, contol = contol) 63 # p = NLP(f, x0, name = 'bench2', df = df, c=c, dc = dc, lb = lb, ub = ub, gtol=gtol, ftol = ftol, maxFunEvals = 1e7, maxIter = 1e4, maxTime = maxTime, plot = 1, color = color, iprint = 0, legend = [solvers[j]], show=False, contol = contol) 64 64 if solver[:4] == ['ralg']: 65 65 pass 66 # p.g radtol = 1e-866 # p.gtol = 1e-8 67 67 # p.ftol = 1e-7 68 68 # p.xtol = 1e-7 PythonPackages/OpenOpt/openopt/tests/nlpLC.py
r170 r748 59 59 60 60 # If you use solver algencan, NB! - it ignores xtol and ftol; using maxTime, maxCPUTime, maxIter, maxFunEvals, fEnough is recommended. 61 # Note that in algencan g radtol means norm of projected gradient of the Augmented Lagrangian61 # Note that in algencan gtol means norm of projected gradient of the Augmented Lagrangian 62 62 # so it should be something like 1e-3...1e-5 63 63 gtol = 1e-7 # (default gtol = 1e-6)
