## Optimisation strategies |

Open the window "Control data for optimisation" by clicking on the <Optimisation> button.
Warning: Diagnosis of input data Check your input data on possible errors or mistakes via Diagnosis before your first opotimisation runs. The outcome of the optimisation primarily depends on complete and error-free data input.
You can use this window to choose between simpler (and faster) and more advanced (and slower) optimisation strategies. You can also select the number of different timetables which will be created during the optimisation run.
The optimisation variants are ordered by:
•optimisation duration (A is the shortest and E is the longest strategy) •recommended work steps (begin with strategy A and work step by step through to the more advanced strategies)
Strategy A - fast optimisation This is the fastest optimisation variant. It does not return the best results but is ideally suited to discovering errors in the input data. For this reason, use this strategy at the start of scheduling until major data errors in master data and lessons have been remedied.
Tip: Errors in input data prevent good optimisation results. Work with strategy A in order to find errors with the input data under 'Diagnosis'.
Strategy B - advanced optimisation This strategy returns very good results and does not take too long. Run this variant after running strategy A and look at the results. If necessary, adjust the weighting slide controls if the timetables do not match your expectations.
Tip: Developing weighting parameters There is an enormous difference between weighting values of 4 and 5. A weighting value of 5 will impair the result even though a value of 4 would be sufficient. For this reason you are advised to set the slide controls to a maximum of 4 and only to change individual controls to 5 one by one if the timetables generated do not match your expectations.
Strategy D - advanced percentage planning Strategies D or B will return the best results depending on the school. However, since strategy D takes considerably longer to run, you should not invoke it until you have developed weighting parameters using strategy B. In this variant, the algorithm proceeds step by step, i.e. does not process all lessons in one go. For this reason you must enter the start and incremental percentage value in the optimisation dialogue.
Tip: Start and incremental percentage value We recommend a starting percentage value of 30% and incremental percentage value of 20%.
Strategy E - overnight optimisation As the name implies, this strategy can take a long time to run, but in most cases returns the best results. Use it at the very end of optimisation, i.e. after using the other strategies. How long this optimisation takes to run depends very much on the size of the school, the number of timetables to be computed, the optimisation steps per timetable and on how powerful the computer is that you are using. It can therefore very well take the whole night to run.
Optimisation steps per timetable (1-9) A value between 1 (low optimisation) and 9 (advanced optimisation) is possible. You can compare the optimisation level to the settings of a chess computer which determine how long the computer may take to analyse each move.
Launching optimisation You can start the optimisation by clickinh on the <OK> button. |