Version 1.9.2

From Heureka Wiki
Jump to: navigation, search

2012-10-16

About version 1.9.2

Version 1.9.2 has several improvements and fixes of the treatment program generator in PlanWise.

You should uninstall previous version before installing version 1.9.2!

Use path heureka.ad.slu.se instead of heureka.resgeom.slu.se. So download from here: Download_and_install

New features

  • Version 1.9.2 has an improved control for linking generation 1 and generation 2 control categories in a forest domain. With the new dialog, you can select explicitly what generation 2 category or categories to link to each generation 1 category. See Template:HeurekeHelpLinkkontrollkategori.htm, section "Koppla kontrollkategori till skogsdomän i PlanVis".
  • Support has been added for a fifth optimization solver, MOSEK, so now there is support for LP_Solve, CPLEX 12.2, Gurobi 5.0, SCIP/Soplex 2.1/3.0 and MOSEK 6.0. MOSEK is very efficient, but a bit cheaper than CPLEX and Gurobi. It is free for academic use. You can get a license from MOSEK's website: [1]

Bug fixes

  • When using breeded plants (förädlade plantor), the treatment program generation in PlanWise could fail in finding treatment programs with thinnings. This has been fixed, but a drawback is longer execution times. To decrease execution times, consider setting "Breeding Improvement Time Horizon" in control table Treatment Program Generator to a smaller value than 100 years, say 50. This means that the production increase factor for plantations made in year 50 would be used in later plantations made after year 50 as well.
  • Selection fellings were not simulated in the last planning period when running treatment program generator in PlanWise.
  • In a treatment unit (stand) with several sample plots (using FMPP data), in which some plot was bare land but the other plots established forest, the treatment program generator could give wrong results when Thinning Decision Level was set to PlotLevel.
  • Trees with a diameter dbh < Min Diameter in Final Fellings (control table Treatment Model), have been ignored, but they were supposed to be part of the soil carbon input pool as harvest residues. This has been fixed. Note that in clear cuttings (without seed trees), thinning grade is reported as 100 %, without consideration to if small trees (dbh < min diameter) are cut or not.
  • In reports over treatment data results or financial results, such as volume harvested, and when at the same time a numeric-type conditions has been used from treatment data (such as "volume harvested where dgv of harvested trees is at least 20 cm"), results were wrong (larger values with than without condition!). This has now been fixed.
  • Result variables belonging to categories "StandObjectData" and "TreatmentUnit" were not available as condition variables in the report builder. This has been fixed.
  • Minor edits have been made to cost functions for harvesting and forwarding, to take into account that trees can be cut for bio fuel instead of timber or pulpwood. If top diameter is changed to a larger value a larger proportion of a tree is available for bio fuel extraction (if this is activated), and forwarding cost should be adjusted accordingly, which is was not. A new result variable has been added under Financial Value, calls Stems Cut Timber & Pulp, for the number of trees which are at least partly cut into timber or pulpwood logs. The difference between TreatmentData.StemsHarvested and FinancialData.StemsCutPulpAndTimber is the number of small trees left after final felling (trees with dbh < minimum diameter).
  • Saving a map view as a jpeg-image now works as it should.
  • Treatment Proposal handling introduced in 1.9.1 had some problem which have now been solved. Especially, planting immediately and regeneration did not work as expected.
  • Swedish language was not available when installing with MSI packages.
  • In StandWise, you could not select Final Felling with Seed Trees (fröträd).
  • In StandWise in Swedish language mode, in the input box for tree species in "Register tree data" dialog, the textual code for Spruce had been mixed up with Pinus mugo (code 11 vs 21). This was a cosmetic bug since the wrong code was displayed immediately. It only occured when entering data manually.
  • A problem when calculating net present value in StandWise that could lead to an infinite loop has been fixed.

Other changes that are not really bugs

  • Variables from category SiteData is now available as condition variables in reports and optimization models.
  • Column-wise table view in result details under Simulation/Optimization tab has been disablad, because it prevented displaying of category variables (such as EvenAged class). Row-wise view, which was already default, shows one row for each treatment unit, period and treatment, while column-wise view showed one row for each treatment unit with all period values in separate columns.
  • When running a simulation on different computers, slightly different results were obtained if the data import had been done locally (to each computer´s own SQL Server database). This was due to a random component invoked when calculating occurrence of damages on sapling that affect growth. Unique identity keys created when importing data were there used as random seeds, hence why results differed. The problem has been fixed by a fixed seed for the random generator as default (this seed can be set in control table Production Model).

Known Issues

see Known Issues