What is OptimiserStarterTargetResourcesPerCluster parameter?

In similar previous incident, it mentioned as below.

OptimiserStarterTargetResourcesPerCluster  value is used to stop the infinite loop in the clustering algorithm.
Ex: if we set OptimiserStarterTargetResourcesPerCluster = 500 so that the clustering algorithm does not get into an infinite loop.

hope this info helps.

Best Regards

-Kelum

I see the default value as 50 in PSO parameters. Are you saying that the lower the value, the clustering algorithm go for infinite loop?

For example, if we set 400, will the clustering algorithm go for infinite loop?

How this paraeter value is going to impact the optimization and the DSE utilization/performance?

Hi @ERCJAGATH 
I have no idea which value is best suited for this.
The idea of that parameter is to stop the infinite loop in the clustering algorithm.
PSO expert will answer your query...

Hi Jagath,

There is something called “Starter Algorithms” which kicks in by DSE to cluster the scheduling problem. Clustering means, DSE should segregate the scheduling problem to smaller chunks. When the size of this chunk is larger, performance requirement of the DSE will also be larger.

Now this parameter determines the number of Resources that should be included in a chunk or cluster when Starter Algorithms are running. This can be increased or decreased with RND recommendation based on the scheduling problem. Now in the issue which Kelum mentioned, there were only few unallocated activities and more than 250 Resources so this parameter has been increased to 500 as a workaround to a bug customer was facing. Strain on the DSE was not an issue in that case.

But your issue is different and you have around 25,000 activities in the schedule. Then you need to reduce the strain on DSE so the option would be to reduce parameter value. That is why RnD has suggested to bring it down to 5 from Default 50.

Regards,
Sajith

Hi

If we decrease the param value from 50 to 5, it may impact/delay the assignment as well?

Thanks,

Jagath