Re: [Kb-complexity] Re: Kb-complexity Digest, Vol 8, Issue 2

[kk aw <kkaw@multicentric.com>]

Victor,

I think the use of the terms "linear" and "non-linear" are misleading.  
It relates to the difficulty of solving non-linear simultaneous 
equations. The real world complex systems may or may not be linear or 
non-linear in parts.  What is important to me is that the system is self 
adaptive.  There is always an element of uncertainty. We are dealing 
with hundreds or thousands of parameters.  There is also massive 
redundancies.  Worst still, they may not follow rules. (e.g. the al 
Qaeda Network) At the moment, a network model appears to be the most 
appropriate, even if the interactions between the various nodes in the 
network are difficult to predict.

With fuzzy logic systems may be dealing with a complex system but each 
component may be linear.  For example, to control the atmosphere in a 
room, you can change the temperature, the air flow or the humidity.  A 
fuzzy logic system will attempt to modify the atmosphere by adjusting 
all three parameters simultaneously.  Based on the feedback and desired 
state, the logic system makes the necessary corrections.  Compared to 
real world problems, this is still a trivial problem.

Climate change issues are obviously very complex.  Depending on the 
assumptions made, one can come to very different conclusions.  The 
change can be both oscillating and at the same time persistent.  Hence, 
depending on the time scale the measurement is made and recorded, we may 
not detect the persistent change.  Worst still, there may be periods of 
calm that distort the picture completely. 

Back to the original question.  Do we need maths to understand 
complexity?  My answer is yes, but you need much more than mathematics.

Cheers,
KK Aw


Victor@34sp wrote:
> Maybe my language wasn't precise enough. As you state a linear system 
> is one in which we can predict its behaviour from a linear equation, 
> e.g. F=ma Using this equation alone can have enormous implications on 
> the real world in terms of being able to predict events that will 
> happen in the future. I tend to even see equations like F=GMM/r2, 
> which is not strictly linear because is has a quadratic term as linear 
> because it is easily tractable. These equations, won't help, however 
> when we try to measure non-linear systems. Even Ed Lotrenz' equations 
> used in his weather forecasting about forty years ago, though they are 
> simple, because terms depend on terms that relate back to themselves, 
> mean the outcomes are far less obvious and even the slightest 
> variation in measurement will predict a significantly different 
> outcome.As you say, biological, ecological and social systems are 
> non-linear and linear methods of trying to measure them and use those 
> results to predict the future are unlikely to provide any real 
> predictability. However, so many people try to measure complex systems 
> using linear measuring tools. As I stated, even making the first step 
> of recognising that non-linear systems need to investigated 
> differently and thinking about non-linear systems from a complexity 
> perspective is in itself a signfiicant step forward.
>
> Cheers
> victor