Process Modelling Optimisation - Have I missed the point here

Hi Folks

Im interested in supply chain optimisation and scheduling along the lines of discrete event analysis.

I love the fact that nupic is great at pattern recognition and anomally detection but how can you get it to determine / recommend optimisations in a supply chain.



1 Like

I moved this from #htm-theory:neuroscience into #other-topics:off-topic because I don’t really know where to put this question.

Hi Craig, I don’t know anything about supply chain optimization (except what I just skimmed on Wikipedia). If you are interested in event analysis, there may be an application there if the events are coming at a fast enough rate.

Can you identify any temporal data streams in the system you want to analyze?

Supply Chain optimisation is normally done with Discrete Event Analysis, see for an overview.

So events don’t typically come in fast as in a single instance / actor within a system however a large system having a lot of moving parts will generate a lot of events. I suppose this qualifies for temporal data stream.

So from a temporal data stream can you optimise / model system resources or constraints?

It is really hard to say if HTM at this point in its evolution would be applicable to this area without an idea about what problems need to be solved and what the data stream actually contains. :confused:

The data streams of the entire events can not be solved. But we can solve the data temporal data streams of exceptional machines and Unite it to form a super computer mind. And we need to start executing it as fast as possible. Not bcoz I am not an optimist and I don’t believe we can change this world. But because I am pessimist who adores the concept of life and would like to find out how wastly a species can push it’s intelligence as a responsible social being. And I just don’t believe we are the answer to every problem. Art and designing is that solution.

This reminds me of some of the reinforced learning problems in the YouTube course by David Sliver (for example, the problem in Lecture 3 at RL Course by David Silver - Lecture 3: Planning by Dynamic Programming - YouTube around time stamp 36:25)

NuPic doesn’t currently do RL, but I agree with @rhyolight, some more information about the problems that need to be solved and some examples of the data stream would help to determine where HTM might be a good fit.

I think @gekosurf is looking for a slightly shorter-term solution :smile:

And I think some super computer implanted the ideas that grew in your brain. They were designing life during the evolution. Alan Turing could view no difference between man and machine and he invented the first machine to do something he was pationate about even though it couldn’t kill it with complex signals. I can visualise it in my virtual memory that contains this streaming data. I have some thing to speak to the programming souls out there