Build 21.4.2 has now been released (13/01/12).
This build has been tested in-house, however we cannot guarantee it to be problem free.
Please trial the new system before committing to it.
NOTE: To install this release you must be running version 21.3.12 or higher, if you are not already running that you must first upgrade to 21.3.12 before you can upgrade to 21.4.2. Version 21.3.12 can be downloaded from the download section of the website.
The major changes in this release are:-
- This issue was a long time coming and that is because it incorporates a major re-write of some of the most critical parts of the scheduler. The changes should manifest in better schedules. You should see an improvement both in time-scales (shorter) and costs (cheaper). The schedule logs now provide more feedback on how effective the new merging system is. The merging system is responsible for detecting and coalescing common sub-assemblies and bought parts. This is, by far, the most complex component of Match-IT. It has undergone extensive testing internally and we think it works as intended, but please be vigilant at first and let us know if you notice anything ‘odd’.
- As part of the above, a new scheduling mode has been added, its called ‘JIT/SPLIT’. In this mode, first a JIT schedule is attempted (Jut-In-Time), if that fails an attempt to break up the demand into chunks that can be achieved is made. For example, say a demand is for 200 for some date and that cannot be achieved, the new mode may discover that it can do 158 on time and the remaining 42 a few days later. In the presence of multiple resource choices the result can be astounding – e.g. if there are 4 machines available the job may split into 4 lots of 50, all on-time.
- The List & Label report generator has also been updated to the latest version (v17). This version contains a number of features that are worth investigating. For example, it is now possible to directly print PDF files that are referenced via Match-IT fields (such as qualifiers). They can even be columns in tables.
- Another significant internal change is the memory allocator. A new much more efficient memory manager is used. You won’t notice much difference until you start doing something requiring ‘heavy’ memory usage, e.g. scheduling very large datasets, or importing large quantities of data via a Lua script, then the difference can be dramatic.