$Id: whats_new_08.omh 3686 2015-05-11 16:14:25Z bradbell $
/* --------------------------------------------------------------------------
CppAD: C++ Algorithmic Differentiation: Copyright (C) 2003-15 Bradley M. Bell

CppAD is distributed under multiple licenses. This distribution is under
the terms of the
                    Eclipse Public License Version 1.0.

A copy of this license is included in the COPYING file of this distribution.
Please visit http://www.coin-or.org/CppAD/ for information on other licenses.
-------------------------------------------------------------------------- */

$begin whats_new_08$$
$spell
	dir
	inline
	tanh
	erf
	nlp
	fg_vector
	Bool
	Ipopt
	Jacobians
	Jacobian
	cygwin
	retape
	det
	Adolc
	gcc
	speedtest
	tarballs
	cppad
	cpl
	gpl
	tgz
	Microsoft
	op
	CppAD
	yum
	vec
	eval
	const
$$

$section Changes and Additions to CppAD During 2008$$


$head Introduction$$
This section contains a list of the changes to CppAD during 2008
(in reverse order by date).
The purpose of this section is to
assist you in learning about changes between various versions of CppAD.

$head 12-19$$
In the documentation for $cref pow_int$$ change the integer exponent
from $codei%int %y%$$ to $codei%const int &%y%$$.
In the implementation for $cref pow$$ make the integer base case
agree with the documentation; i.e., change
from $codei%int %x%$$ to $codei%const int &%x%$$.

$head 12-14$$
Added another test of $cref mul_level$$ calculations (in the
$code test_more$$ directory).

$head 12-04$$
Extensive explanation for the $code ipopt_cppad/ipopt_cppad_ode$$ example was
provided in the section $cref/ipopt_cppad_ode/ipopt_nlp_ode/$$.

$head 11-22$$
The CppAD interface to the Ipopt nonlinear programming solver
$cref cppad_ipopt_nlp$$
has been moved from $code example/ipopt_cppad_nlp$$ to
$code ipopt_cppad/ipopt_cppad_nlp$$.

$head 11-21$$
The Microsoft's Visual C++ Version 9.0 generates a warning of the form
$codei warning C4396:%...%$$
for every template function that is declared as a both a friend and inline
(it thinks it is only doing this for specializations of template functions).
The warnings are no longer generated because
these $code inline$$ directives are converted to
empty code when a Microsoft Visual C++ is used.


$head 11-20$$
$index erf$$
The function $codei%tanh(%x%)%$$ was added to the
$cref unary_standard_math$$ functions.
The $code abs$$ and $code erf$$ functions were removed from the
$cref/Base requirements/base_require/$$.
The restrictions about the Base class were removed from
$cref abs$$, $cref atan2$$, $cref LuRatio$$, $cref erf$$.
$pre

$$
Visual Studio Version 9.0 could not handle the large number of static
constants in the CppAD $cref erf$$ function.
This function was changed to a simpler representation
that is much faster and that
is differentiable at all points (not defined differently on subregions).
The down side to this is that the new version
is not as accurate.

$head 10-27$$
Change prototypes for $code ipopt_cppad/ipopt_cppad_ode$$ helper routines
to use $code const$$ (where appropriate).

$head 10-17$$
Major improvements to the $code ipopt_cppad/ipopt_cppad_ode$$ example.

$head 10-16$$
Minor improvement to description of optimization argument in
$code ipopt_cppad/ipopt_cppad_ode$$.

$head 09-30$$
Add or modify some wish list entries; see
$code cppad_ipopt_nlp$$ (since removed),
multiple argument forward (completed with $cref forward_dir$$),
and sparsity patterns
($cref/sparsity patterns/glossary/Sparsity Pattern/$$ has been fulfilled).


$head 09-26$$
Use parenthesis and brackets to group terms of the form
$latex m \times I$$ to make the documentation of
$cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ easier to read.

Changed $code ipopt_cppad/ipopt_cppad_ode$$ to use $latex y(t)$$ for
the solution of the ODE to distinguish it for $latex x$$,
the vector we are optimizing with respect to.

$head 09-18$$
Changed $code ipopt_cppad/ipopt_cppad_ode$$ to a case where $latex x(t)$$
is a pair of exponential functions instead of a linear and quadratic.
Fixed some of the comments in this example and included the source
code in the documentation (which was missing by mistake).

$head 09-17$$
Changed $code ipopt_cppad/ipopt_cppad_ode$$ to a case where there are
two components in the ODE (instead of one).
Also removed an initialization section that was only intended for tests
with a specific initial value.

$head 09-16$$
Add $code ipopt_cppad/ipopt_cppad_ode$$, an example and test
that optimizes the solution of an ODE.
Change $code r_eval$$ to $code eval_r$$ in $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$.
Fix a dimension of $code u_ad$$ error in $code ipopt_cppad_nlp$$.

$head 09-12$$
Converted from storing full Hessian and Jacobian to a sparse data structure
in $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$.
This greatly reduced the memory requirements (and increased the speed)
for sparse problems.

$head 09-10$$
Fixed more indexing bugs in $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$
that effected cases where the domain index vector $latex J_{k, \ell}$$
was different for different values of $latex k$$ and $latex \ell$$.
$pre

$$
In $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$,
combined $icode%fg_info%->domain_index()%$$ and
$icode%fg_info%->range_index()%$$ into a single function called
$icode%fg_info%->index()%$$.
Also added more error checking (if $code NDEBUG$$ is not defined).

$head 09-09$$
Fixed an indexing bug in $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$.
(This effected cases where the domain index vector $latex J_{k, \ell}$$
was different for different values of $latex k$$ and $latex \ell$$.)

$head 09-07$$
Change $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ so that object and constraints
are expressed as the double summation of simpler functions.
This is more versatile that the single summation representation.

$head 09-06$$
Checked in a major change to $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$
whereby the object and constraints can be expressed as the
sum of simpler functions.
This is the first step in what will eventually be a more versatile
representation.

$head 09-05$$
Fix bug in $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ (not recording the function
at the proper location. Here is the difference that occurred
multiple places in the $code ipopt_cppad/ipopt_cppad_nlp.cpp$$ source:
$codep
	for(j = 0; j < n_; j++)
-		x_ad_vec[0] = x[j];
+		x_ad_vec[j] = x[j];
$$
This did not show up in testing because there currently is no test of
$code ipopt_cppad_nlp$$ where the operation sequence depends
on the value of $latex x$$.
$pre

$$
Changed $code eval_grad_f$$ in $code ipopt_cppad_nlp.cpp$$
to be more efficient.

$head 09-04$$
The $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ interface has been changed to use
a derived class object instead of a pointer to a function.

$head 09-03$$
The $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ interface has been changed to use
$code size_t$$ instead of $code Ipopt::Index$$.

$head 09-01$$
Back out the changes made to $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ on 08-29
(because testing proved the change to be less efficient in the case
that motivated the change).

$head 08-29$$
The $code push_vector$$ member function was missing from the
$cref/vectorBool/Cppad_vector/vectorBool/$$ class.
This has been fixed.
In addition, it seems that for some cases (or compilers)
the assignment
$codei%
	%x%[%i%] = %y%[%j%]
%$$
did not work properly when both $icode x$$ and $icode y$$
had type $code vectorBool$$.
This has been fixed.
$pre

$$
$index ipopt_cppad_nlp$$
The $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ example has been extended
so that it allows for both scalar and vector evaluation of the
objective and constraints; see the argument $icode fg_vector$$ in
$cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$.
In the case where there is not a lot of common terms
between the functions, the scalar evaluation may be more efficient.

$head 08-19$$
Add $cref/push of a vector/CppAD_vector/push_vector/$$
to the $code CppAD::vector$$ template class.
This makes it easy to accumulate multiple scalars and
$cref/simple vectors/SimpleVector/$$ into one large $code CppAD::vector$$.

$head 08-08$$
There was an indexing bug in the
$cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$ example
that affected the $icode retape$$ equal to $code false$$ case.
This has been fixed.
In addition, the missing
$icode retape$$ documentation was added.

$head 07-02$$
Extend $cref/configure/auto_tools/Configure/$$ command to check for
extras libraries that are necessary for linking the ipopt example.

$head 06-18$$
Add specifications for the Ipopt class $cref/ipopt_cppad_nlp/cppad_ipopt_nlp/$$.
This is only an example class it may change with future
versions of CppAD.

$head 06-15$$
The nonlinear programming example $cref ipopt_nlp_get_started.cpp$$ was added.
This is a preliminary version of this example.

$head 06-11$$
The sparsity pattern for the Hessian was being calculated each time
by $cref/SparseHessian/sparse_hessian/$$.
This is not efficient
when the pattern does not change between calls to $code SparseHessian$$.
An optional sparsity pattern argument was added to $code SparseHessian$$
so that it need not be recalculated each time.

$head 06-10$$
The sparsity pattern for the Jacobian was being calculated each time
by $cref/SparseJacobian/sparse_jacobian/$$.
This is not efficient
when the pattern does not change between calls to $code SparseJacobian$$.
An optional sparsity pattern argument was added to $code SparseJacobian$$
so that it need not be recalculated each time.

$head 05-08$$
The $cref sparse_jacobian$$ routine has been added.
$pre

$$
The example in $cref sparse_hessian$$ pointed to $cref hessian.cpp$$
instead of $cref sparse_hessian.cpp$$.
This has been fixed.

$head 05-03$$
The $code retape$$ flag has been added to $cref speed_main$$.
In addition the routines
$cref link_det_minor$$, $cref link_poly$$, and $cref link_ode$$
pass this flag along to the speed test implementations
(because the corresponding tests have a fixed operation sequence).
If this flag is false, a test implementation is allowed to just tape
the operation sequence once and reuse it.
The following tests use this flag:
$cref adolc_det_minor.cpp$$,
$cref cppad_det_minor.cpp$$,
$cref cppad_ode.cpp$$,
$cref adolc_poly.cpp$$,
$cref cppad_poly.cpp$$.
$pre

$$
Create specialized zero order forward mode routine that should be faster,
but does not test out as faster under cygwin g++ (GCC) 3.4.4.




$head 04-20$$
Added the $cref ode_evaluate$$ speed test utility in
preparation for having ode speed tests.
Created ode speed test for the $code cppad$$ and $code double$$
cases; see $cref speed_main$$.
In addition,
added the examples $cref ode_evaluate.cpp$$ and $cref sparse_hessian.cpp$$.
$pre

$$
Changed the $cref speed_main$$ routines defined for each package
from $codei%compute_%name%$$ to $codei%link_%name%$$.
For example, in $code speed/cppad/det_minor.cpp$$,
the function name
$code compute_det_minor$$ was changed to $code link_det_minor$$.


$head 04-18$$
Fix a problem in the $cref link_poly$$ correctness test.
Also add $cref double_sparse_hessian.cpp$$ to the set
speed and correctness tests (now available).


$head 04-10$$
Change all the
$cref/Adolc speed/speed_adolc/$$ examples to use
$cref TrackNewDel$$ instead of using $code new$$ and $code delete$$
directly.
This makes it easy to check for memory allocation errors and leaks
(when $code NDEBUG$$ is not defined).
Also include in documentation sub functions that indicate
the $code sparse_hessian$$ speed test is not available for
$cref double_sparse_hessian.cpp$$,
$cref fadbad_sparse_hessian.cpp$$, and
$cref sacado_sparse_hessian.cpp$$.

$head 04-06$$
The following $cref/wish list/WishList/$$
entry has been completed and removed from the list:
"Change private member variables names
(not part of the user interface)
so that they all end with an underscore."

$head 04-04$$
Fix a problem compiling the speed test
$cref/main/speed_main/$$ program with gcc 4.3.

$head 03-27$$
Corrected $cref cppad_sparse_hessian.cpp$$
so that it uses the sparse case when
$code USE_CPPAD_SPARSE_HESSIAN$$ is $code 1$$.
Also added a wish list sparsity pattern entry
(the $cref/sparsity pattern/glossary/Sparsity Pattern/$$ entry
has been fulfilled).
$pre

$$
Change the name of $code speedtest.cpp$$ to $cref speed_program.cpp$$.


$head 02-05$$
Change windows install instructions
to use Unix formatted files (so only two instead of
four tarballs are necessary for each version).
The Microsoft project files for
$code speed/cppad$$, $code speed/double$$, and $code speed/example$$
were missing. This has also been fixed.

$head 02-03$$
There was an ambiguity problem
(detected by g++ 4.3) with the following operations
$codei%
	%x% %op% %y%
%$$
where $icode x$$ and $icode y$$ were $code AD<double>$$ and
$icode op$$ was a member operator of that class.
This has been fixed by making all such member functions friends
instead of members of $code AD<double>$$.
$pre

$$
Remove computed assignment entry from wish list
(it was fixed on $cref/2007-05-26/whats_new_07/05-26/$$).
Add
$cref/expression hashing/WishList/Optimization/Expression Hashing/$$
to the wish list.
Add Library and Scripting Languages to the wish list
(this has since been fulfilled by the example $cref ad_in_c.cpp$$).

$head 01-26$$
The $cref LuFactor$$ routine gave a misleading error message
when the input matrix had not a number or infinity in it.
This has been fixed.

$head 01-24$$
The $cref/postfix_dir/auto_tools/postfix_dir/$$
has been added to the $code configure$$ command line options.

$head 01-21$$
A sparse Hessian case was added to the $cref speed$$ tests; see
$cref/sparse_hessian/link_sparse_hessian/$$.

$head 01-20$$
CppAD can now be installed using $code yum$$ on
$code Fedora$$ operating systems.


$head 01-11$$
The CppAD correctness tests assume that machine epsilon is less
than $code 1e-13$$. A test for this has been added to the
$code test_more/test_more$$ program.

$head 01-08$$
Added a $cref sparse_hessian$$ routine and
extended $cref Hessian$$ to allow for a weight vector $icode w$$
instead of just one component $icode l$$.

$end