- Boost - lists.boost.org

Re: [boost] Proposal for a Differential Evolution C++ library
by Adrian Michel 18 Jan '12

18 Jan '12

Hi, Sorry for the late reply. On 1/10/2012 11:01 AM, Topher Cooper wrote: > > On 1/10/2012 1:37 AM, Adrian Michel wrote: >> The way I see it, the only common point between different optimization >> algorithms (if they are generic enough) is the objective function, >> which should have a very simple prototype: >> >> double f( std::vector< double> ); >> >> so in theory if different algorithms are implemented to accept the >> same types (in C++ sense) of objective functions, then you could run >> them interchangeably between GA, DE etc without any modification to >> the function or algorithm. > > That ignores multi-objective optimization. In that case the return type > would be, at a minimum std:vector<double>, and might require information > on what characteristics are maximized. > ... > The points you raised make sense, but I think they go beyond the intended scope of this initial release. Unlike many other boost libraries whose main purpose is to enhance the language, DE is intended almost exclusively for solving domain specific problems. Because DE can be applied in many areas, by developers with a variety of requirements and levels of knowledge, I believe taking an iterative approach based on users' feedback will increase the chances of creating a library that's useful in solving real problems. As I mentioned before, this is an initial implementation. Some of the features I intend to add in a later release are: - multi-objective optimization - constraint functions (as opposed to just plain boundary constraints) - design centering ("Design centering is a constraint satisfaction problem whose goal is to find the solution, which, when perturbed, remains feasible more often than any other solution") and others. But these features are not trivial, and it would help me to get a sense of how the library is used before I start working on them. So far, I have received several good suggestions, I would like to nail down a list of specific additional requirements that this library needs to meet before moving on to the next step (and I'm not sure what that may be). Please let me know how to proceed. Adrian www.tradery.com

1 0

Re: [boost] Proposal for a Differential Evolution C++library
by Adrian Michel 18 Jan '12

18 Jan '12

On 1/10/2012 1:32 AM, Graham, Jeff wrote: > > I can help test on i686, amd64 and arm9 running linux. > Great, thanks! Can you please download the library from http://www.amichel.com/de/de.zip and try to build it? Let me know if you need help or have any questions. Best, Adrian www.tradery.com

1 0

Re: [boost] Proposal for a Differential Evolution C++ library
by Adrian Michel 18 Jan '12

18 Jan '12

On 1/11/2012 11:29 AM, Simonson, Lucanus J wrote: >The DE library description needs to be > very clear that it is for non-linear numerical optimization. I'm also > curious about how it compares industrial solvers like CPLEX. That is what I > had in mind when I said "good". It may be DE is one of the best ways to > solve a class of problems for which we don't know of a good way. There is > no point arguing about it in subjective terms, however. Instead, > quantifying how big of a problem it can handle in how many compute hours and > then providing a reference comparison to other solvers for other classes of > problem will put it into context and people will understand easily that it > is what it is and be able to tell if it is right for their problem or not. > There are people who may need it and not know it. There are people who may > need something else and not know it. The documentation should help both types of people. > Noted. In general I prefer to link to other sources that deal with the subject in much greater detail, but I can certainly add more information to help first time readers get a sense of why this library may be useful to them. > As far as design goes I would say that you want to split it into generic > genetic algorithm engine and a module for performing DE mutations. The GA > and DE-mutation modules would be configured separately and then composed to > produce the solution. That means you need to define an interface between > the GA and DE modules, but once you do that then you are able to easily swap > out GA engines or apply different mutation modules to achieve different GA > algorithms. I don't think that the mutation module needs to couple to > anything to do with the objective function at all. If someone wants to plug > in their HADOOP powered genetic algorithm instead of using the library > provided one then they would just use the DE mutations module. Alternately > they could use ASIO to build their own distributed GA module based on the > one provided by the library. There are already several policy components to allow user customization of DE, one of them being the "Mutation Policy" component. Does this address your suggestion, or are you referring to something even more general than that? Thanks, Adrian www.tradery.com

1 0

[config] How to use on Mac with C++11 dialect?
by John M. Dlugosz 17 Jan '12

17 Jan '12

I've asked this on Boost-User and had no reply. Maybe someone here would know? Using the current Boost 1.48 and the current Clang for the Apple Macintosh, when the compiler is set to accept the C++11 dialect (that is, new "auto" usage, lambda functions, rvalue references, etc.) then Boost headers choke. So currently we can use either C++11 or Boost, but not both. I'm sure it's a matter of configuration, of telling Boost headers that the compiler is indeed supporting these constructs. Can someone please point out what needs to be done? Thanks, —John

3 2

Re: [boost] compile time parser generator
by Martin Bidlingmaier 16 Jan '12

16 Jan '12

I'm not sure it's possible to get a string literal in a template< char ... > class string; 'Literals transformation is redefined into two distinct phases: raw and cooked. A raw literal is a sequence of characters of some specific type, while the cooked literal is of a separate type. The C++ literal 1234, as a raw literal, is this sequence of characters '1', '2', '3', '4'. As a cooked literal, it is the integer 1234. The C++ literal 0xA in raw form is '0', 'x', 'A', while in cooked form it is the integer 10. Literals can be extended in both raw and cooked forms, with the exception of string literals, which can only be processed in cooked form. This exception is due to the fact that strings have prefixes that affect the specific meaning and type of the characters in question.' - source: wikipedia - http://en.wikipedia.org/wiki/C%2B%2B11#User-defined_literals . Parsing string literals with constexpr is described in this article: http://akrzemi1.wordpress.com/2011/05/11/parsing-strings-at-compile-time-pa… For building the parser generator I used compile time polymorphism. Each symbol in EBNF would be represented as a class. The interface each symbol defines looks like this: struct symbol { constexpr symbol( const_string str , unsigned int begin = 0 ); //for an implementation of const_string see the article I linked constexpr bool valid(); constexpr unsigned int end(); }; For example, a symbol that parses a single character 'a' looks like this: struct sym_a { constexpr sym_a( const_string str , unsigned int begin = 0 ) : m_end ( begin >= str.size() ? 0u - 1 : //invalid index str[ begin ] == 'a' ? begin + 1 : 0u - 1 ) { } constexpr bool valid() { return m_end != 0u - 1; } constexpr unsigned int end() { return m_end; } unsigned int m_end; }; constexpr sym_a s1( "a" ); constexpr sym_a s2( "b" ); static_assert( s1.valid() , "" ); static_assert( !s2.valid() , "" ); The source for my current version of the parser generator is appended. I've also created a simple calculator that can handle + and * (and will accept any digit sequence as number, including "00000"). Notice, that currently only char strings are completely supported, defining terminals and the opt< T > operator won't work with other literals. (I know I can improve this for terminals, but I'm not sure about opt< T >.) The source compiles with gcc 4.6 and c++11 turned on without errors. -- NEU: FreePhone - 0ct/min Handyspartarif mit Geld-zurück-Garantie! Jetzt informieren: http://www.gmx.net/de/go/freephone

6 19

Re: [boost] boost threads: pthread mutex priority inheritance + patches
by Jason Newton 16 Jan '12

16 Jan '12

Ack, no discussion on this since I posted it back in July. In my opensuse packages, I apply the following patch to insure priority inheritence on mutexs. I had to go inside the mutex ctor to do the work. -Jason On Wed, Jul 27, 2011 at 5:40 PM, Jason Newton <nevion(a)gmail.com> wrote: > On Wed, Jul 27, 2011 at 4:48 PM, Jason Newton <nevion(a)gmail.com> wrote: > >> For the moment, the few lines that compose the mutex class are easily >> copied so that I can do my own business inside the constructor but I expect >> others will run into this problem down the line and this solution is a bit >> unsavory. Any better advice? >> > > Actually a private copy of mutex didn't work so well overall as pthread > conditions are hard coded to > void boost::condition_variable::wait(boost::unique_lock<boost::mutex>&) > > -Jason >

2 1

"protected" APIs
by Lorenzo Caminiti 16 Jan '12

16 Jan '12

Hello all, Boost.Closure (formerly, Boost.Local) needs to use some macros, types, etc that are currently marked private in Boost.ScopeExit: BOOST_SCOPE_EXIT_AUX_TAG boost::scope_exit::aux::undeclared ... So I will move these symbols from AUX/aux to another prefix/namespace. How shall the new prefix/namespace be named? These symbols should not go into the public API of ScopeExit because they should never be used by the end user. These symbols are only meant to be used by other parts of Boost so they should use some sort of "protected" prefix/namespace. For example, I could use "detail": "aux" for ScopeExit private API and "detail" for ScopeExit API that can be used within Boost but not by general users: BOOST_SCOPE_EXIT_DETAIL_TAG boost::scope_exit::detail::undeclared However, other Boost components already use "detail" for their private API (synonymous of "aux"). So, I could use "protected" instead: BOOST_SCOPE_EXIT_PROTECTED_TAG boost::scope_exit::protected_::undeclared What's the best convention to follow here? Thanks a lot. --Lorenzo

8 8

build error in Release branch
by Barend Gehrels 15 Jan '12

15 Jan '12

Hi, I get a build error in the release branch: .../release/tools/build/v2/build/project.jam:266: unbalanced parentheses I just looked for differences in Trunk and Release and indeed that line differs. I updated the file conform trunk, locally, and I can build now. I then get the message: warning: mismatched versions of Boost.Build engine and core warning: Boost.Build engine (bjam) is 2011.12.00 warning: Boost.Build core (at /home/barend/svn/boost/release/tools/build/v2) is 2011.04-svn So this might be my problem (not sure). However I'm normally using only one bjam, copying it to /usr/local/bin, but this time I cannot do that... Should I use different bjam's for Trunk/Release versions? Is this a known problem / will this be solved? Regards, Barend

2 2

[scope_exit] improvements
by Lorenzo Caminiti 15 Jan '12

15 Jan '12

Hello all, I've been working with Alex to make the following improvements to Boost.ScopeExit (as they were discussed during Boost.Local's review). The update docs: https://svn.boost.org/svn/boost/sandbox/closure/libs/scope_exit/doc/html/in… 1) It is possible to use variadic macros on compilers that support them: int r = 0, v = 0; BOOST_SCOPE_EXIT(&r, v) { r = v; } BOOST_SCOPE_EXIT_END To support compilers without variadics and for backward compatibility with previous ScopeExit versions, all macros also accepts pp-sequences: BOOST_SCOPE_EXIT( (&r) (v) ) { r = v; } BOOST_SCOPE_EXIT_END 2) It is possible to capture no variable using `void`: BOOST_SCOPE_EXIT(void) { std::cout << "cout is a global variable << std::endl; } BOOST_SCOPE_EXIT_END This same syntax works on compilers with and without variadics. 3) From within member functions, it is possible to capture the object this using `this_`: BOOST_SCOPE_EXIT(&r, v, this_) { r = this_->f(v); } BOOST_SCOPE_EXIT_END On C++11, you can also (but you don't have to) use `this`: BOOST_SCOPE_EXIT(&r, v, this) { // C++11 only. r = this->f(v); } BOOST_SCOPE_EXIT_END 4) On C++11, you can (but you don't have to) use `;` instead of BOOST_SCOPE_END. 5) On C++11, there is an additional macro that allows to capture all variables using lambda-style `&` or `=`: BOOST_SCOPE_EXIT_ALL(&, v, this) { // C++11 only. r = this->f(v); }; Or equivalently: BOOST_SCOPE_EXIT_ALL(=, &r) { // C++11 only. r = this->f(v); }; I still have to move these changes from the sandbox into trunk. Comments are welcome. Thanks, --Lorenzo

2 1

[typeof] TYPEOF(this) on MSVC 7.1 and 8.0
by Lorenzo Caminiti 14 Jan '12

14 Jan '12

Hello all, BOOST_TYPEOF(this) generates error C2355 on MSVC 7.1 and 8.0 because of a MSVC bug. The code below was suggested to workaround the problem but it only works with native typeof. Is there a way to make BOOST_TYPEOF(this) work on MSVC 7.1 and 8.0 with Boost.Typeof in emulation mode? // Adam Butcher's workaround to deduce `this` type on MSVC revision < 10. // Boost.Typeof for VC71's typeid-based workaround does not work to determine // `this` type due to error C2355 being incorrectly reported. The typical // avoidance strategy implemented below is to make an indirect compile-time // constant by assigning an enum and use that as type-index-- this only works // with the sizeof() approach and not with the typeid() approach. This does not // work in typeof emulation mode. #if BOOST_WORKAROUND(BOOST_MSVC, >= 1300) && /* MSVC 7.1 */ \ BOOST_WORKAROUND(BOOST_MSVC, <= 1400) && /* MSVC 8.0 */ \ !defined(BOOST_TYPEOF_EMULATION) /* doesn't work in typeof-emu mode */ namespace boost { namespace scope_exit { namespace aux { namespace msvc_typeof_this { template<int Id> struct msvc_typeid_wrapper { typedef typename boost::type_of::msvc_extract_type<boost::mpl::int_<Id> >::id2type id2type; typedef typename id2type::type type; }; template<> struct msvc_typeid_wrapper<4> { typedef msvc_typeid_wrapper<4> type; }; template<typename T> struct encode_type { BOOST_STATIC_CONSTANT(unsigned, value = boost::type_of::encode_counter<T>::count); typedef typename boost::type_of::msvc_register_type<T, boost::mpl::int_<value> >::id2type type; BOOST_STATIC_CONSTANT(unsigned, next = value + 1); }; template<class T> struct sizer { typedef char(*type)[encode_type<T>::value]; }; template<typename T> typename boost::disable_if< typename boost::is_function<T>::type , typename sizer<T>::type >::type encode_start(T const&); template<typename T> typename boost::enable_if< typename boost::is_function<T>::type , typename sizer<T>::type >::type encode_start(T&); template<typename Organizer, typename T> boost::type_of::msvc_register_type<T, Organizer> typeof_register_type(const T&, Organizer* = 0); } } } } #define BOOST_SCOPE_EXIT_AUX_TYPEDEF_TYPEOF_THIS_INDEX_(id) \ BOOST_PP_CAT(boost_se_thistype_index_, id) #define BOOST_SCOPE_EXIT_DETAIL_TYPEDEF_TYPEOF_THIS(id, new_type) \ enum { \ BOOST_SCOPE_EXIT_AUX_TYPEDEF_TYPEOF_THIS_INDEX_(id) = sizeof( \ *::boost::scope_exit::aux::msvc_typeof_this::encode_start(this)) \ }; \ typedef ::boost::scope_exit::aux::msvc_typeof_this::msvc_typeid_wrapper< \ BOOST_SCOPE_EXIT_AUX_TYPEDEF_TYPEOF_THIS_INDEX_(id) \ >::type new_type; #else // typeof(this) on other compilers are OK. #define BOOST_SCOPE_EXIT_DETAIL_TYPEDEF_TYPEOF_THIS(id, new_type) \ typedef BOOST_TYPEOF(this) new_type; #endif // typeof(this) For example: // within a member function BOOST_SCOPE_EXIT_DETAIL_TYPEDEF_TYPEOF_THIS(__LINE__, this_type); this_type obj = this; Thanks. --Lorenzo

1 0