From 9944c256e0b893ac79decfaf7d7937933bd203d5 Mon Sep 17 00:00:00 2001
From: Erin Kuci <erinkuci@gmail.com>
Date: Wed, 25 Nov 2020 12:23:52 +0100
Subject: [PATCH] add continuation on mma parameter for constraint relaxation
---
Ccore/onelab_optimize.py | 8 ++++-
Lbracket/topo.pro | 9 +++---
Lbracket/topo.py | 65 +++++++++++++++++++++++++++++++---------
Team25/shape.py | 7 +++++
4 files changed, 70 insertions(+), 19 deletions(-)
diff --git a/Ccore/onelab_optimize.py b/Ccore/onelab_optimize.py
index 6833da8..b353d7a 100644
--- a/Ccore/onelab_optimize.py
+++ b/Ccore/onelab_optimize.py
@@ -130,7 +130,13 @@ def Optimize(Optimizer):
conveks.mma.initialize(initialPoint, lowerBound, upperBound)
# Set some options for MMA
- conveks.mma.option.setNumber('General.Verbosity', 0) # def 4
+ if Optimizer == 1:
+ conveks.mma.option.setNumber('General.Verbosity', 0) # def 4
+ conveks.mma.option.setNumber('SubProblem.isContinuation', 1)
+ conveks.mma.option.setNumber('SubProblem.cInitial' , 1e2)
+ conveks.mma.option.setNumber('SubProblem.cFinal' , 1e5)
+ conveks.mma.option.setNumber('SubProblem.cMultiply' , 10)
+ conveks.mma.option.setNumber('SubProblem.cIteration', 50)
# Get iteration count (here it will be 1 - could be different in case of restart)
it = conveks.mma.getOuterIteration()
diff --git a/Lbracket/topo.pro b/Lbracket/topo.pro
index 9aa2aea..895f076 100644
--- a/Lbracket/topo.pro
+++ b/Lbracket/topo.pro
@@ -9,6 +9,7 @@ DefineConstant [
Optimizer = {0,Choices {0="Conveks - MMA",1="Conveks - GCMMA"},Name "Optimization/00Optimizer"}
Opt_maxIter = {1600, Name "Optimization/01Max iterations"}
Opt_maxChange = {0.01, Name "Optimization/02Max change"}
+ Opt_isInnerIter = {0, Name "Optimization/03Inner Iteration", Visible 0, Choices {0, 1}}
// Density field parameters
densityFieldInit = {0.5, Name "Optimization/3Density/0Inital value"}
@@ -624,16 +625,16 @@ PostOperation {
Format Table, StoreInVariable $vonMises_qp_pnorm, File "", Color "LightYellow",
SendToServer StrCat[Opt_ResDir_Onelab,"pnorm of Von-Mises"] ] ;
- If(Flag_PrintLevel > 2)
+ If(Flag_PrintLevel > 2 && !Opt_isInnerIter)
CreateDir[Opt_ResDir];
Print[ vonMises_qp, OnElementsOf Vol_Elast~{iP},
File StrCat[Opt_ResDir,"vm_qp.pos"]];
EndIf
- If(Flag_PrintLevel > 4)
+ If(Flag_PrintLevel > 4 && !Opt_isInnerIter)
Print[ vonMises, OnElementsOf Vol_Elast~{iP},
File StrCat[Opt_ResDir,"vm.pos"]];
EndIf
- If(Flag_PrintLevel > 5)
+ If(Flag_PrintLevel > 5 && !Opt_isInnerIter)
Print[ Young, OnElementsOf Vol_Elast~{iP},
File StrCat[Opt_ResDir, "young.pos"] ];
//Print[ Felast_pressure, OnElementsOf Sur_Force~{iP},
@@ -762,7 +763,7 @@ Resolution {
InitSolution[LAM];
// Show the density field
- If(Flag_PrintLevel > 0)
+ If(Flag_PrintLevel > 0 && !Opt_isInnerIter)
PostOperation[Get_DensityField];
EndIf
diff --git a/Lbracket/topo.py b/Lbracket/topo.py
index bc333d1..d4375c0 100644
--- a/Lbracket/topo.py
+++ b/Lbracket/topo.py
@@ -131,10 +131,18 @@ conveks.initialize()
# Initialize the MMA optimizer
conveks.mma.initialize(x, lowerBound, upperBound)
-
# Set appropriate options for MMA
conveks.mma.option.setNumber('General.Verbosity', 0)
-if Optimizer==0:conveks.mma.option.setNumber('SubProblem.move', 0.1)
+# if Optimizer==0:conveks.mma.option.setNumber('SubProblem.move', 0.1)
+# conveks.mma.option.setNumber('SubProblem.asymptotesRmax', 10)
+# conveks.mma.option.setNumber('SubProblem.asymptotesRmin', 0.01)
+# conveks.mma.option.setNumber('SubProblem.adaptSubproblem', 1)
+if Optimizer != 1:
+ conveks.mma.option.setNumber('SubProblem.isContinuation', 1)
+ conveks.mma.option.setNumber('SubProblem.cInitial' , 1e2)
+ conveks.mma.option.setNumber('SubProblem.cFinal' , 1e5)
+ conveks.mma.option.setNumber('SubProblem.cMultiply' , 10)
+ conveks.mma.option.setNumber('SubProblem.cIteration', 50)
# Get iteration count (here it will be 1 - could be different in case of restart)
it = conveks.mma.getOuterIteration()
@@ -154,17 +162,28 @@ c.setNumber('Optimization/Results/objective', value=objective)
c.addNumberChoice('Optimization/Results/objective', value=objective)
c.setNumber('Optimization/Results/max(|Constraints|)', value=np.max(np.abs(constraints)))
c.addNumberChoice('Optimization/Results/max(|Constraints|)', value=np.max(np.abs(constraints)))
+c.setNumber('Optimization/03Inner Iteration', value=0)
+show_inner = 1
# show current iteration
+# print 'iter. inner iter. obj. max(constr.) L2-norm(kkt) point'
+# print '%3i %3i %2.4e %2.4e %2.4e'%(it-1, innerit, objective, np.max(np.abs(constraints)), kkt_norm),
+# print x[0:5]
print 'iter. inner iter. obj. max(constr.) L2-norm(kkt) point'
-print '%3i %3i %2.4e %2.4e %2.4e'%(it-1, innerit, objective, np.max(np.abs(constraints)), kkt_norm),
-print x[0:5]
+print '%3i %3i %.6e %.6e %.6e %3i'%(it-1, innerit, objective, np.max(np.abs(constraints)), kkt_norm, 0),
+print '[',
+for xk in x[0:5]:
+ print '%.6e'%(float(xk)),
+print ']'
while change > maxChange and it <= maxIter and c.getString('topo/Action') != 'stop':
# Solve MMA
conveks.mma.updateCurrentPoint(objective,grad_objective,constraints,grad_constraints)
+ # and get the total number of nonlinear loops
+ nlloops = conveks.mma.countNonlinearIterations()
+
# Let us check now if the MMA approximation is conservative
# at the new point; and adapt the latter if it is not the case.
if Optimizer == 1:
@@ -174,29 +193,42 @@ while change > maxChange and it <= maxIter and c.getString('topo/Action') != 'st
conserv = conveks.mma.isConservativeSubProblem(obj_new, c_new)
innerit = 0
- print '\t it. new-point (xn) f(xc) f(xn) max(c(xc)) max(c(xn))'
- print '\t%3i'%(innerit),
- print conveks.mma.getCurrentPoint()[0:5],
- print ' %.4e %.4e %.4e %.4e'%(objective,obj_new,np.max(constraints), np.max(c_new))
+ # print '\t it. new-point (xn) f(xc) f(xn) max(c(xc)) max(c(xn))'
+ # print '\t%3i'%(innerit),
+ # print conveks.mma.getCurrentPoint()[0:5],
+ # print ' %.4e %.4e %.4e %.4e'%(objective,obj_new,np.max(constraints), np.max(c_new))
while innerit <= 15 and conserv==0:
+ # Do not show fields (density, von-Mises, ...) during inner iterations
+ c.setNumber('Optimization/03Inner Iteration', value=1)
innerit += 1
# adapt the subproblem to make it conservative and generate a new point
conveks.mma.updateCurrentPoint(objective,grad_objective,
objectiveAtUpdatedPoint=obj_new,constraintsAtUpdatedPoint=c_new)
+ # and get the total number of nonlinear loops
+ nlloops = conveks.mma.countNonlinearIterations()
+
# evaluate the objective and the constraints at the new point
# and check if the subproblem is conservative at this point
obj_new,c_new = get_objective(conveks.mma.getCurrentPoint())
conserv = conveks.mma.isConservativeSubProblem(obj_new, c_new)
# show inner iteration
- print '\t%3i'%(innerit),
- print conveks.mma.getCurrentPoint()[0:5],
- print ' %.4e %.4e %.4e %.4e'%(objective,obj_new,np.max(constraints), np.max(c_new))
+ # print '\t%3i'%(innerit),
+ # print conveks.mma.getCurrentPoint()[0:5],
+ # print ' %.4e %.4e %.4e %.4e %3i'%(objective,obj_new,np.max(constraints), np.max(c_new), nlloops)
+ if show_inner:
+ print '%3i %3i '%(it, innerit),
+ print '%.6e %.6e %.6e %3i'%(obj_new,np.max(np.abs(c_new)), kkt_norm, nlloops),
+ print '[',
+ for xk in conveks.mma.getCurrentPoint()[0:5]:
+ print '%.6e'%(float(xk)),
+ print ']'
inneritTot += innerit
+ c.setNumber('Optimization/03Inner Iteration', value=0)
# get (copy of) current point
x = conveks.mma.getCurrentPoint()
@@ -208,10 +240,15 @@ while change > maxChange and it <= maxIter and c.getString('topo/Action') != 'st
# evaluate the L2-norm of KKT
kkt_norm = conveks.mma.getKKTNorm(constraints,grad_objective,grad_constraints)
-
+
# show the current state
- print '%3i %3i %2.4e %2.4e %2.4e'%(it, innerit, objective, np.max(np.abs(constraints)), kkt_norm),
- print x[0:5]
+ # print '%3i %3i %2.4e %2.4e %2.4e %4i'%(it, innerit, objective, np.max(np.abs(constraints)), kkt_norm, nlloops),
+ # print x[0:5]
+ print '%3i %3i %.6e %.6e %.6e %3i'%(it, innerit, objective, np.max(np.abs(constraints)), kkt_norm, nlloops),
+ print '[',
+ for xk in x[0:5]:
+ print '%.6e'%(float(xk)),
+ print ']'
# get the outer iteration count here as well as the
# change between two successive points
diff --git a/Team25/shape.py b/Team25/shape.py
index 60b3200..53d4bb1 100644
--- a/Team25/shape.py
+++ b/Team25/shape.py
@@ -155,6 +155,13 @@ conveks.mma.initialize(initialPoint, lowerBound, upperBound)
# Set some options for MMA
conveks.mma.option.setNumber('General.Verbosity', 0*4)
+if Optimizer != 1:
+ conveks.mma.option.setNumber('General.Verbosity', 0) # def 4
+ conveks.mma.option.setNumber('SubProblem.isContinuation', 1)
+ conveks.mma.option.setNumber('SubProblem.cInitial' , 1e2)
+ conveks.mma.option.setNumber('SubProblem.cFinal' , 1e5)
+ conveks.mma.option.setNumber('SubProblem.cMultiply' , 10)
+ conveks.mma.option.setNumber('SubProblem.cIteration', 50)
# Get iteration count (here it will be 1 - could be different in case of restart)
it = conveks.mma.getOuterIteration()
--
GitLab