Fixed the compilation bug having been identified on Fedora 14: https://bugzilla.redhat.com/show_bug.cgi?id=631080
The logs of the Fedora 14 build are accessible here:
RMOL has been officially accepted to be distributed by Fedora. RPM packages are available for Fedora, on the official site, from Fedora version 10 (e.g., http://fr2.rpmfind.net/linux/fedora/updates/10/x86_64/rmol-0.19.0-1.fc10.i386.rpm).
On a Fedora system, if the fedora-updates repository is enabled (which is the case by default), you have therefore just to type the following command to be able to integrate the RMOL library within your own piece of software:
yum -y install rmol-devel
The new release, namely the version 0.17.0, brings, along the classical platform-independent tar-balls (g-zipped and Bzip2-ed), RPM packages. Those RPM packages have been generated on a Linux Fedora 10 distribution, but may work on most of the RPM-based distributions (e.g., Red Hat, Mandriva, Novell Suse). Moreover, the source RPM file (rmol.*.src.rpm) should allow to build the RPM packages tailored to your specific distribution, if needed.
jRMOM is the Java version of the RMOL library. Why Java? For an even wider spread of RMOL of course!
As the RPM package has been built on a Fedora Core 6 with an AMD64, I do not know whether other platforms and/or processors are supported as well.
From the source tarball and the spec file, however, it is possible to build a RPM on a Suse platform (I have tested it successfully on a Suse 9.2 with AMD64).
So, if some of you want to add RPM support for other platform and/or processor, do not hesitate to upload such files.
The building system now follows the GNU standards (GNITS, etc.). Hence, the classical "./configure', 'make' and 'make install' sequence now works.
Note that the test has to be compiled with the check target: 'make check' (there's not even the need to install the library for that purpose).
The documentation has been borrowed from another project, and the content must be changed.
The i18n has been borrowed to still another project, and the content must also be changed.
That first working version optimally calculates the protections for the buckets/classes, corresponding to the prices/yields and demand distribution parameters (mean and standard deviation) given as input (as of today, still hard-coded in the "main" source code file, i.e., optimise.cpp).
The algorithm is a slight variation of the "Monte Carlo Integration" algorithm, given by Kalyan T. Talluri and Garret J. van Ryzin in their book "The Theory and Practice of Revenue Management" (Kluwer Academic Publishers) pp 42-43.... read more