I used quasi-newton method to locate the minimum. It used lnsrch subroutine instead of linmin.
But every time, the subroutine returns with the condition "f.le.fold+ALF*alam*slope", but not "alam .lt. alamin".
I don't quite understand what "f.le.fold+ALF*alam*slope" really means.
And whether it is correct or not when it returns here?

This means that the function has decreased sufficiently with that value of lambda, which is a good thing in terms of getting the method to converge.

I have some doubts....suppose i need to find the max of a function then what changes in need to do in dfpmin and lnsrch .....i already tried the same code by multiplying negative to the max function to make it min function.......but couldn't get any proper answer...please help....