Maintaining SQ Build Strategy progress through reboots

I asked this question to support in an email but thought I would post it here for everyone’s benefit as well…

I have a question about how to have SQ “continue” building strategies where it left off prior to a reboot instead of starting all over.

I asked this question to support in an email but thought I would post it here for everyone’s benefit as well…

I have a question about how to have SQ “continue” building strategies where it left off prior to a reboot instead of starting all over.

During a Building Strategies session using the Genetic method, I can see that the app “remembers” whether it has created and evaluated a certain strategy already, and therefore can retain it or dismiss it per the entries in the log. Of course this is very useful to allow the app to be efficient in the build strategy process.

However, am I correct in assuming that the “memory” of the app only lasts during that specific build strategy session. So if the current session is stopped and then another build strategy is started using exactly the same settings as the previous session, then the app may well tread the same path and generate thousands of strategies that are precisely the same as in the previous session?

For example, let’s assume I have two instances of SQ on my machine and have allocated 2 cores and 4 GB of RAM to each one. On Instance 01, I build strategies for 24 hours. After 24 hours, I stop session #1 after it has built 2500 strategies that meet my ranking options. I save all of those strategies to a file and prepare to Test them. To allow me to make the most of my time, I intend to use Instance 02 SQ to do the retests. This allows me to put Instance 01 back to work generating more strategies. However, before I start testing or building new strategies, I need to reboot my machine. So I handle that task first.

After the reboot, let’s assume I start up Instance 01 of SQ and use exactly the same settings for build session #2 as I did for build session #1. Does SQ automatically “know” which strategies that it built in session #1 so that it doesn’t rebuild them in session #2?

If not, is there a setting that I can turn on/off so that it will remember and pick up in build session #2 where it left off in build session #1?

Thanks in advance.

Odoslané z G3121 pomocou Tapatalku

 

I asked this question to support in an email but thought I would post it here for everyone’s benefit as well…

 

I have a question about how to have SQ “continue” building strategies where it left off prior to a reboot instead of starting all over.

 

During a Building Strategies session using the Genetic method, I can see that the app “remembers” whether it has created and evaluated a certain strategy already, and therefore can retain it or dismiss it per the entries in the log. Of course this is very useful to allow the app to be efficient in the build strategy process. 

 

However, am I correct in assuming that the “memory” of the app only lasts during that specific build strategy session. So if the current session is stopped and then another build strategy is started using exactly the same settings as the previous session, then the app may well tread the same path and generate thousands of strategies that are precisely the same as in the previous session?

 

For example, let’s assume I have two instances of SQ on my machine and have allocated 2 cores and 4 GB of RAM to each one. On Instance 01, I build strategies for 24 hours. After 24 hours, I stop session #1 after it has built 2500 strategies that meet my ranking options. I save all of those strategies to a file and prepare to Test them. To allow me to make the most of my time, I intend to use Instance 02 SQ to do the retests. This allows me to put Instance 01 back to work generating more strategies. However, before I start testing or building new strategies, I need to reboot my machine. So I handle that task first.

 

After the reboot, let’s assume I start up Instance 01 of SQ and use exactly the same settings for build session #2 as I did for build session #1. Does SQ automatically “know” which strategies that it built in session #1 so that it doesn’t rebuild them in session #2?

 

If not, is there a setting that I can turn on/off so that it will remember and pick up in build session #2 where it left off in build session #1?

 

Thanks in advance.

 

Kornel responded to me in another venue and this is what he shared with me:

“StrategyQuant start every generating with initial random set of strategies. So even if you will run exactly the same setting at two machines, then finded strategies will always differ. Currently Strategyquant don’t have function to save the memory.”

So this being the case, in theory, I assume the best way to generate strategies is to let the build process continue running as long as you possibly can and just save off and delete strategies out of the Databank when you have a substantial set you want to test and let the app fill it up again. As long as the app continues to run, it should be able to identify strategies it has already evaluated. 

Of course, this means that the longer it runs the longer it has to hold more strategies in memory. The impact on each user’s machine will be different so that has to be taken into consideration.