New changes from l2g
w
This commit is contained in:
@@ -3,11 +3,14 @@
|
||||
implicit none
|
||||
include '../testarea/LeafGasParams.h'
|
||||
include '../testarea/LeafGasHybridFit.h'
|
||||
integer i,ndim,k,j,iderivative,iwrong
|
||||
integer i,ndim,k,j,iderivative,iwrong,n,icompete,i2,isitnaninf,
|
||||
&nave
|
||||
double precision beta(4),sumsquare0,beta0(4),sumsquarecp,
|
||||
&betacp(4),ftol,xtol,shortx(maxobs,2),shorty(maxobs),
|
||||
&xvar(maxobs,2),weitx(maxobs,2),weity(maxobs),ran2,
|
||||
&templflights0(maxobs),aparlights0(maxobs),termmin,termmax
|
||||
&xvar(maxobs,2),weitx(maxobs,2),weity(maxobs),
|
||||
&templflights0(maxobs),aparlights0(maxobs),termmin,termmax,
|
||||
&ftol_relax,term1,term2,ran2,discount,history(2000,10),upper,lower,
|
||||
&f1dim_flujmax,flujmax_pikaia
|
||||
parameter(ftol=1.0d-7,xtol=1.0d-7)
|
||||
external funkmin_flujmax,f1dim_flujmax,FCN_flujmax,flujmax_pikaia
|
||||
!beta(1)=fjmax25
|
||||
@@ -65,104 +68,298 @@
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
history(1,i)=beta(i)
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
history(1,ndim+1)=flujmaxfval
|
||||
!entrance counter
|
||||
history(1,ndim+2)=1.0d0
|
||||
!failure counter
|
||||
history(1,ndim+3)=0.0d0
|
||||
!Is it a competition among different initial guesses?
|
||||
icompete=0
|
||||
!j the total number of calls to nongradopt; k is the number of returns to the current best and reset
|
||||
!to zero if a better minumum is found; n is the number of scouting points over the landscape of the cost function.
|
||||
!The first initial guess provided by the user is always part of the set of scouting points.the rest consist of outcomes
|
||||
!from calls to nongradopt if they are significantly different from the current best.
|
||||
j=0
|
||||
k=0
|
||||
30 call nongradopt(ndim,funkmin_flujmax,
|
||||
&f1dim_flujmax,beta,betamin,betamax,ftol,flujmaxfval)
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
if((flujmaxfval+1.0d0).eq.flujmaxfval)then
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
else
|
||||
if(dabs(flujmaxfval-sumsquare0).lt.ftol)k=k+1
|
||||
if(flujmaxfval.gt.sumsquare0)then
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
else
|
||||
if((sumsquare0-flujmaxfval).gt.ftol)k=0
|
||||
!reset the counter of revisiting a minimum if a new minimum is found
|
||||
endif
|
||||
endif
|
||||
j=j+1
|
||||
!try different initial guesses
|
||||
if(j.lt.200.and.k.lt.50)then
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
beta(i)=betamin(i)+ran2()*(betamax(i)-betamin(i))
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
goto 30
|
||||
endif
|
||||
call RepeatCompassSearch(ndim,beta,flujmaxfval,betamin,
|
||||
&betamax,funkmin_flujmax,f1dim_flujmax,xtol)
|
||||
do i=1,ndim
|
||||
betacp(i)=(beta(i)-betamin(i))/(betamax(i)-betamin(i))
|
||||
n=1
|
||||
nave=1
|
||||
ftol_relax=ftol*1000.0d0
|
||||
discount=2.0d0
|
||||
30 do i=1,ndim
|
||||
betacp(i)=beta(i)
|
||||
enddo
|
||||
isitbounded=0
|
||||
call pikaia(flujmax_pikaia,ndim,gacontrol,betacp,flujmaxfval,i)
|
||||
do i=1,ndim
|
||||
beta(i)=betamin(i)+betacp(i)*(betamax(i)-betamin(i))
|
||||
enddo
|
||||
isitbounded=1
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
call RepeatCompassSearch(ndim,beta,flujmaxfval,betamin,
|
||||
&betamax,funkmin_flujmax,f1dim_flujmax,xtol)
|
||||
isitbounded=1
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
sumsquarecp=flujmaxfval
|
||||
iderivative=0
|
||||
iwrong=0
|
||||
call odr_leastsquare(ndim,FCN_flujmax,beta,ntotlights,
|
||||
&xvar(1:ntotlights,1:2),2,flphips2lights,1,weitx(1:ntotlights,1:2),
|
||||
&weity,iderivative,shortx(1:ntotlights,1:2),shorty(1:ntotlights),
|
||||
&flujmaxfval,iwrong)
|
||||
isitbounded=1
|
||||
!after odr_leastsquare, forcing variables are destroyed. restore to the origninals
|
||||
do i=1,ntotlights
|
||||
templflights(i)=templflights0(i)
|
||||
aparlights(i)=aparlights0(i)
|
||||
enddo
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
if(dabs(flujmaxfval).le.dabs(sumsquare0))then
|
||||
if(isitnaninf(flujmaxfval).eq.1.or.flujmaxfval.gt.sumsquarecp)then
|
||||
do i=1,ndim
|
||||
beta(i)=betacp(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquarecp
|
||||
else
|
||||
if(dabs(flujmaxfval).gt.1.0d+20)then
|
||||
!in case of infinity (division by zero)
|
||||
do i=1,ndim
|
||||
betacp(i)=beta(i)
|
||||
enddo
|
||||
sumsquarecp=flujmaxfval
|
||||
endif
|
||||
call nongradopt(ndim,funkmin_flujmax,f1dim_flujmax,
|
||||
&beta,betamin,betamax,ftol_relax,flujmaxfval)
|
||||
if(isitnaninf(flujmaxfval).eq.1.or.flujmaxfval.gt.sumsquarecp)then
|
||||
do i=1,ndim
|
||||
beta(i)=betacp(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquarecp
|
||||
endif
|
||||
if(flujmaxfval.gt.1.0d0)then
|
||||
term1=flujmaxfval*ftol_relax
|
||||
else
|
||||
term1=ftol_relax*10.0d0
|
||||
endif
|
||||
if(flujmaxfval.gt.sumsquare0)then
|
||||
!failure
|
||||
if((flujmaxfval-sumsquare0).gt.term1)then
|
||||
if(icompete.eq.1)history(1,ndim+3)=history(1,ndim+3)+1.5d0
|
||||
!even though flujmaxfval is much worse than sumsquare0, it is an output of optimization after all so
|
||||
!include it in the set if it has not already been included in the set.
|
||||
i=1
|
||||
i2=1
|
||||
40 if(dabs(history(i2,i)-beta(i)).gt.ftol_relax)then
|
||||
if(dabs(history(i2,ndim+1)-flujmaxfval).lt.term1)then
|
||||
history(i2,ndim+3)=history(i2,ndim+3)+1.0d0
|
||||
goto 60
|
||||
endif
|
||||
if(i2.ge.n)goto 50
|
||||
i2=i2+1
|
||||
i=1
|
||||
goto 40
|
||||
else
|
||||
if(i.ge.ndim)goto 60
|
||||
i=i+1
|
||||
goto 40
|
||||
endif
|
||||
50 n=n+1
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
history(n,i)=beta(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
history(n,ndim+1)=flujmaxfval
|
||||
history(n,ndim+2)=0.0d0
|
||||
history(n,ndim+3)=0.0d0
|
||||
!use average only when there is improvement
|
||||
nave=n
|
||||
else
|
||||
!designed this way to avoid flujmaxfval='NAN'
|
||||
!the difference is minimal even though flujmaxfval is larger than sumsquare0.
|
||||
!Increment the counter for arriving at the same minimum.
|
||||
if(icompete.eq.1)history(1,ndim+3)=history(1,ndim+3)+1.0d0
|
||||
k=k+1
|
||||
endif
|
||||
60 do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
else
|
||||
!success
|
||||
if((sumsquare0-flujmaxfval).lt.term1)then
|
||||
!negligible improvement. Increment the counter for arriving at the same minimum.
|
||||
!no increment for the set of central initial guesses
|
||||
if(icompete.eq.1)history(1,ndim+3)=history(1,ndim+3)+0.5d0
|
||||
k=k+1
|
||||
else
|
||||
!reset the counter for arriving at a better minimum.
|
||||
!Increment the set of central initial guesses
|
||||
if(dabs(discount-2.0d0).lt.ftol)then
|
||||
discount=dmax1(0.001d0,(sumsquare0-flujmaxfval)/1000.0d0)
|
||||
endif
|
||||
k=0
|
||||
n=n+1
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
history(n,i)=beta(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
history(n,ndim+1)=flujmaxfval
|
||||
history(n,ndim+2)=0.0d0
|
||||
history(n,ndim+3)=0.0d0
|
||||
endif
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
endif
|
||||
j=j+1
|
||||
if(j.lt.200.and.k.lt.3)then
|
||||
!first explore around the very first initial guess
|
||||
if(j.lt.10)then
|
||||
term1=0.05d0+dmin1(history(1,ndim+3)*0.1d0,0.9d0)
|
||||
history(1,ndim+2)=history(1,ndim+2)+1.0d0
|
||||
do i=1,ndim
|
||||
lower=history(1,i)-term1*(history(1,i)-betamin(i))
|
||||
upper=history(1,i)+term1*(betamax(i)-history(1,i))
|
||||
beta(i)=lower+ran2()*(upper-lower)
|
||||
enddo
|
||||
icompete=1
|
||||
goto 70
|
||||
endif
|
||||
!try average
|
||||
if(n.gt.nave)then
|
||||
term1=1.0d0/(history(1,ndim+1)+1.0d-5)
|
||||
do i=2,n
|
||||
term1=term1+1.0d0/(history(i,ndim+1)+1.0d-5)
|
||||
enddo
|
||||
do i=1,ndim
|
||||
beta(i)=history(1,i)/(term1*(history(1,ndim+1)+1.0d-5))
|
||||
do icompete=2,n
|
||||
beta(i)=beta(i)+history(icompete,i)/
|
||||
&(term1*(history(icompete,ndim+1)+1.0d-5))
|
||||
enddo
|
||||
enddo
|
||||
nave=n
|
||||
icompete=0
|
||||
goto 70
|
||||
endif
|
||||
!try different initial guesses
|
||||
if(ran2().gt.0.2d0)then
|
||||
!guess around the best
|
||||
icompete=1
|
||||
term1=history(1,ndim+1)+
|
||||
&discount*history(1,ndim+2)*history(1,ndim+3)
|
||||
do i=2,n
|
||||
term2=history(i,ndim+1)+
|
||||
&discount*history(i,ndim+2)*history(i,ndim+3)
|
||||
if(term2.le.term1)then
|
||||
term1=term2
|
||||
do i2=1,ndim+3
|
||||
history(n+1,i2)=history(i,i2)
|
||||
history(i,i2)=history(1,i2)
|
||||
history(1,i2)=history(n+1,i2)
|
||||
enddo
|
||||
endif
|
||||
enddo
|
||||
term1=0.05d0+dmin1(history(1,ndim+2)*history(1,ndim+3)*
|
||||
&0.015d0,0.9d0)
|
||||
history(1,ndim+2)=history(1,ndim+2)+1.0d0
|
||||
do i=1,ndim
|
||||
lower=history(1,i)-term1*(history(1,i)-betamin(i))
|
||||
upper=history(1,i)+term1*(betamax(i)-history(1,i))
|
||||
beta(i)=lower+ran2()*(upper-lower)
|
||||
enddo
|
||||
else
|
||||
!completely random guess
|
||||
do i=1,ndim
|
||||
beta(i)=betamin(i)+ran2()*(betamax(i)-betamin(i))
|
||||
enddo
|
||||
icompete=0
|
||||
endif
|
||||
70 call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
goto 30
|
||||
else
|
||||
if((ftol_relax-ftol).gt.ftol)then
|
||||
if(k.le.1)then
|
||||
n=n+1
|
||||
do i=1,ndim+3
|
||||
history(n,i)=history(1,i)
|
||||
enddo
|
||||
do i=1,ndim
|
||||
history(1,i)=beta(i)
|
||||
enddo
|
||||
history(1,ndim+1)=flujmaxfval
|
||||
history(1,ndim+2)=0.0d0
|
||||
history(1,ndim+3)=0.0d0
|
||||
do i=1,n
|
||||
do icompete=1,ndim
|
||||
betacp(icompete)=history(i,icompete)
|
||||
enddo
|
||||
sumsquarecp=history(i,ndim+1)
|
||||
call RepeatCompassSearch(ndim,betacp,sumsquarecp,
|
||||
&betamin,betamax,funkmin_flujmax,f1dim_flujmax,ftol_relax)
|
||||
call funkmin_flujmax(ndim,betacp,sumsquarecp)
|
||||
if(isitnaninf(sumsquarecp).eq.0.and.sumsquarecp.lt.
|
||||
&flujmaxfval)then
|
||||
do icompete=1,ndim
|
||||
beta(icompete)=betacp(icompete)
|
||||
enddo
|
||||
flujmaxfval=sumsquarecp
|
||||
endif
|
||||
enddo
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
j=0
|
||||
icompete=1
|
||||
else
|
||||
icompete=0
|
||||
endif
|
||||
ftol_relax=ftol
|
||||
goto 30
|
||||
endif
|
||||
endif
|
||||
j=0
|
||||
100 if(j.ge.10)then
|
||||
|
||||
goto 110
|
||||
|
||||
call RepeatCompassSearch(ndim,beta,flujmaxfval,betamin,
|
||||
&betamax,funkmin_flujmax,f1dim_flujmax,xtol)
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
if(isitnaninf(flujmaxfval).eq.1.or.flujmaxfval.gt.sumsquare0)then
|
||||
do i=1,ndim
|
||||
betacp(i)=(beta(i)-betamin(i))/(betamax(i)-betamin(i))
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
isitbounded=0
|
||||
call pikaia(flujmax_pikaia,ndim,gacontrol,betacp,flujmaxfval,i)
|
||||
flujmaxfval=sumsquare0
|
||||
else
|
||||
do i=1,ndim
|
||||
beta(i)=betamin(i)+betacp(i)*(betamax(i)-betamin(i))
|
||||
beta0(i)=beta(i)
|
||||
enddo
|
||||
isitbounded=1
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
sumsquare0=flujmaxfval
|
||||
endif
|
||||
sumsquare0=flujmaxfval
|
||||
do i=1,ndim
|
||||
betacp(i)=(beta(i)-betamin(i))/(betamax(i)-betamin(i))
|
||||
enddo
|
||||
call pikaia(flujmax_pikaia,ndim,gacontrol,betacp,flujmaxfval,i)
|
||||
if(i.eq.0)then
|
||||
do i=1,ndim
|
||||
betacp(i)=betamin(i)+betacp(i)*(betamax(i)-betamin(i))
|
||||
enddo
|
||||
call funkmin_flujmax(ndim,betacp,flujmaxfval)
|
||||
if(isitnaninf(flujmaxfval).eq.0.and.flujmaxfval.lt.sumsquare0)
|
||||
&then
|
||||
do i=1,ndim
|
||||
beta(i)=betacp(i)
|
||||
beta0(i)=betacp(i)
|
||||
enddo
|
||||
sumsquare0=flujmaxfval
|
||||
endif
|
||||
endif
|
||||
flujmaxfval=sumsquare0
|
||||
iderivative=0
|
||||
iwrong=0
|
||||
call odr_leastsquare(ndim,FCN_flujmax,beta,ntotlights,
|
||||
&xvar(1:ntotlights,1:2),2,flphips2lights,1,weitx(1:ntotlights,1:2),
|
||||
&weity,iderivative,shortx(1:ntotlights,1:2),shorty(1:ntotlights),
|
||||
&flujmaxfval,iwrong)
|
||||
!after odr_leastsquare, forcing variables are destroyed. restore to the origninals
|
||||
do i=1,ntotlights
|
||||
templflights(i)=templflights0(i)
|
||||
aparlights(i)=aparlights0(i)
|
||||
enddo
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
if(isitnaninf(flujmaxfval).eq.1.or.flujmaxfval.gt.sumsquare0)then
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
endif
|
||||
j=0
|
||||
100 sumsquare0=flujmaxfval
|
||||
do i=1,ndim
|
||||
beta0(i)=beta(i)
|
||||
enddo
|
||||
@@ -170,21 +367,11 @@
|
||||
&beta,betamin,betamax,ftol,flujmaxfval)
|
||||
call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
if(flujmaxfval.eq.sumsquare0)return
|
||||
if(dabs(flujmaxfval).le.dabs(sumsquare0))then
|
||||
else
|
||||
if(dabs(flujmaxfval).gt.1.0d+20)then
|
||||
!in case of infinity (division by zero)
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
else
|
||||
!designed this way to avoid flujmaxfval='NAN'
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
endif
|
||||
if(isitnaninf(flujmaxfval).eq.1.or.flujmaxfval.gt.sumsquare0)then
|
||||
do i=1,ndim
|
||||
beta(i)=beta0(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquare0
|
||||
endif
|
||||
sumsquarecp=flujmaxfval
|
||||
do i=1,ndim
|
||||
@@ -194,14 +381,15 @@
|
||||
&betamax,funkmin_flujmax,f1dim_flujmax,xtol)
|
||||
call funkmin_flujmax(ndim,betacp,sumsquarecp)
|
||||
if(flujmaxfval.eq.sumsquarecp)return
|
||||
if(dabs(sumsquarecp).lt.dabs(flujmaxfval))then
|
||||
if(isitnaninf(sumsquarecp).eq.0.and.flujmaxfval.gt.sumsquarecp)
|
||||
&then
|
||||
do i=1,ndim
|
||||
beta(i)=betacp(i)
|
||||
enddo
|
||||
flujmaxfval=sumsquarecp
|
||||
endif
|
||||
j=j+1
|
||||
if(j.le.2.and.dabs(flujmaxfval-sumsquare0).gt.ftol)goto 100
|
||||
if(j.le.2.and.(sumsquare0-flujmaxfval).gt.ftol)goto 100
|
||||
!
|
||||
!------------------------------------------------------
|
||||
110 call funkmin_flujmax(ndim,beta,flujmaxfval)
|
||||
|
||||
Reference in New Issue
Block a user