Discussion of article "Parallel Particle Swarm Optimization" - page 3

 

With the tester_file directive, will this file be forwarded to the Agent during a job bundle change?

If yes, it is possible to change the contents of this file during optimisation.

 
Прилагаемую библиотеку PSO можно использовать как в режиме математических вычислений
What is the fastest way to load large amounts of data on each pass?
 
fxsaber #:

With the tester_file directive, will this file be forwarded to the Agent during a job pack change?

If yes, it is possible to change the contents of this file during optimisation.

You can change the content, but it will have no effect - it seems that the information is cached/stored somewhere during the start of optimisation. At least it was so before.

 
fxsaber #:
What is the fastest way to load large amounts of data on each pass?

For local agents it is probably the fastest via a shared directory. For distributed agents there is not much choice.

 
Stanislav Korotky #:

There is little choice for the distributed.

Embed in source ZIP(MqlTick[]).

 
fxsaber #:

Embed in the source ZIP(MqlTick[]).

How will it differ from tester_file or more precisely from resource? Most likely, the access time will differ, but insignificantly.

 
Stanislav Korotky #:

How will it differ from tester_file or more precisely from resource? Most likely, the access time will be different, but insignificantly.

It will not sit in a sandbox and depend on FileLoad speed. You should measure it, in general.