Changeset 771 for mystic

Timestamp:

11/19/14 10:05:33 (18 months ago)

Author:

mmckerns

Message:

save allSolvers and monitors in ensemble solver; lattice solver default nbins=8

Location:

mystic

Files:

• examples/buckshot_example06.py (modified) (1 diff)
• examples/lattice_example06.py (modified) (1 diff)
• mystic/abstract_ensemble_solver.py (modified) (1 diff)
• mystic/ensemble.py (modified) (11 diffs)

mystic/examples/buckshot_example06.py

@@ -75 +75 @@
     random.seed(123)
     ndim = 9
-    npts = 6
+    npts = 8
 
     # draw frame and exact coefficients

mystic/examples/lattice_example06.py

@@ -75 +75 @@
     random.seed(123)
     ndim = 9
-   #nbins = [1]*ndim
-    nbins = [2,1,2,1,2,1,1,1,1]
+    nbins = [2,1,2,1,2,1,2,1,1]
 
     # draw frame and exact coefficients

mystic/mystic/abstract_ensemble_solver.py

@@ -129 +129 @@
         self._bestSolver      = None # 'best' solver (after Solve)
         self._total_evals     = 0 # total function calls (after Solve)
+        if kwds.has_key('nbins'): NP = reduce(lambda x,y:x*y, nbins)
+        else: NP = npts #XXX: merge npts/nbins kwds?
+        self._allSolvers      = [None for j in range(NP)]
         return
 

mystic/mystic/ensemble.py

@@ -135 +135 @@
 
         # run optimizer for each grid point
-        from copy import deepcopy as copy
-        op = [copy(solver) for i in range(len(initial_values))]
+        from copy import deepcopy as _copy
+        op = [_copy(solver) for i in range(len(initial_values))]
        #cf = [cost for i in range(len(initial_values))]
         vb = [verbose for i in range(len(initial_values))]
@@ -144 +144 @@
         # generate the local_optimize function
         def local_optimize(solver, x0, rank=None, disp=False, callback=None):
+            from copy import deepcopy as _copy
+            from mystic.tools import isNull
             solver.id = rank
             solver.SetInitialPoints(x0)
@@ -150 +152 @@
                                        max=solver._strictMax) # or lower,upper ?
             solver.Solve(cost, disp=disp, callback=callback)
-            return solver
+            sm = solver._stepmon
+            em = solver._evalmon
+            if isNull(sm): sm = ([],[],[],[])
+            else: sm = (_copy(sm._x),_copy(sm._y),_copy(sm._id),_copy(sm._info))
+            if isNull(em): em = ([],[],[],[])
+            else: em = (_copy(em._x),_copy(em._y),_copy(em._id),_copy(em._info))
+            return solver, sm, em
 
         # map:: solver = local_optimize(solver, x0, id, verbose)
@@ -158 +166 @@
         # save initial state
         self._AbstractSolver__save_state()
+        #XXX: HACK TO GET CONTENT OF ALL MONITORS
+        # reconnect monitors; save all solvers
+        from mystic.monitors import Monitor
+        while results: #XXX: option to not save allSolvers? skip this and _copy
+            _solver, _stepmon, _evalmon = results.pop()
+            sm = Monitor()
+            sm._x,sm._y,sm._id,sm._info = _stepmon
+            _solver._stepmon.extend(sm)
+            del sm
+            em = Monitor()
+            em._x,em._y,em._id,em._info = _evalmon
+            _solver._evalmon.extend(em)
+            del em
+            self._allSolvers[len(results)] = _solver
+        del results, _solver, _stepmon, _evalmon
+        #XXX: END HACK
+
         # get the results with the lowest energy
-        self._bestSolver = results[0]
+        self._bestSolver = self._allSolvers[0]
         bestpath = self._bestSolver._stepmon
         besteval = self._bestSolver._evalmon
         self._total_evals = self._bestSolver.evaluations
-        for solver in results[1:]:
+        for solver in self._allSolvers[1:]:
             self._total_evals += solver.evaluations # add func evals
             if solver.bestEnergy < self._bestSolver.bestEnergy:
@@ -294 +319 @@
 
         # run optimizer for each grid point
-        from copy import deepcopy as copy
-        op = [copy(solver) for i in range(len(initial_values))]
+        from copy import deepcopy as _copy
+        op = [_copy(solver) for i in range(len(initial_values))]
        #cf = [cost for i in range(len(initial_values))]
         vb = [verbose for i in range(len(initial_values))]
@@ -303 +328 @@
         # generate the local_optimize function
         def local_optimize(solver, x0, rank=None, disp=False, callback=None):
+            from copy import deepcopy as _copy
+            from mystic.tools import isNull
             solver.id = rank
             solver.SetInitialPoints(x0)
@@ -309 +336 @@
                                        max=solver._strictMax) # or lower,upper ?
             solver.Solve(cost, disp=disp, callback=callback)
-            return solver
+            sm = solver._stepmon
+            em = solver._evalmon
+            if isNull(sm): sm = ([],[],[],[])
+            else: sm = (_copy(sm._x),_copy(sm._y),_copy(sm._id),_copy(sm._info))
+            if isNull(em): em = ([],[],[],[])
+            else: em = (_copy(em._x),_copy(em._y),_copy(em._id),_copy(em._info))
+            return solver, sm, em
 
         # map:: solver = local_optimize(solver, x0, id, verbose)
@@ -317 +350 @@
         # save initial state
         self._AbstractSolver__save_state()
+        #XXX: HACK TO GET CONTENT OF ALL MONITORS
+        # reconnect monitors; save all solvers
+        from mystic.monitors import Monitor
+        while results: #XXX: option to not store allSolvers? skip this & _copy
+            _solver, _stepmon, _evalmon = results.pop()
+            sm = Monitor()
+            sm._x,sm._y,sm._id,sm._info = _stepmon
+            _solver._stepmon.extend(sm)
+            del sm
+            em = Monitor()
+            em._x,em._y,em._id,em._info = _evalmon
+            _solver._evalmon.extend(em)
+            del em
+            self._allSolvers[len(results)] = _solver
+        del results, _solver, _stepmon, _evalmon
+        #XXX: END HACK
+
         # get the results with the lowest energy
-        self._bestSolver = results[0]
+        self._bestSolver = self._allSolvers[0]
         bestpath = self._bestSolver._stepmon
         besteval = self._bestSolver._evalmon
         self._total_evals = self._bestSolver.evaluations
-        for solver in results[1:]:
+        for solver in self._allSolvers[1:]:
             self._total_evals += solver.evaluations # add func evals
             if solver.bestEnergy < self._bestSolver.bestEnergy:
@@ -366 +416 @@
 
 
-def lattice(cost,ndim,nbins=None,args=(),bounds=None,ftol=1e-4,maxiter=None, \
+def lattice(cost,ndim,nbins=8,args=(),bounds=None,ftol=1e-4,maxiter=None, \
             maxfun=None,full_output=0,disp=1,retall=0,callback=None,**kwds):
     """Minimize a function using the lattice ensemble solver.
@@ -381 +431 @@
     cost -- the Python function or method to be minimized.
     ndim -- dimensionality of the problem.
-    nbins -- tuple of number of bins in each dimension. [default = (2,)*ndim]
+    nbins -- tuple of number of bins in each dimension. [default = 8 bins total]
 
 Additional Inputs:
@@ -448 +498 @@
         termination = VTRChangeOverGeneration(ftol)
 
-    if nbins is None: nbins = 2**ndim
-
     solver = LatticeSolver(ndim,nbins)
     solver.SetNestedSolver(_solver) #XXX: skip settings for configured solver?