motivation for application of swarm intelligence

[curious_mind]

Recently I was reading about GA and PSO. Wanted to know the exact motivation that lead to GA and PSO. Could you please give some examples, which cannot be solved by conventional method and hence require PSO or GA to do that. (I prefer easy to understand examples, internet provides relatively complex examples difficult to understand)

 

[srizbf]

Recently I was reading about GA and PSO. Wanted to know the exact motivation that lead to GA and PSO. Could you please give some examples, which cannot be solved by conventional method and hence require PSO or GA to do that. (I prefer easy to understand examples, internet provides relatively complex examples difficult to understand)

Find the minimum or maximum from voluminus Data Set
of measurement or observations , which cannot be
represented by equations or functions..

 

[curious_mind]

Could you explain me further? Why can we not find maxima and minima from dataset that you describe?

 

[srizbf]

Suppose there are 'n' blocks or modules .
Each module is connected with other module with varying
connections.
What is the best arrangement of or permutations of 'n'
modules such that their total connections of all modules are minimum?

 

[curious_mind]

Are there any worked out examples to illustrate how traditional methods fail, while GA wins over

 

[srizbf]

Are there any worked out examples to illustrate how traditional methods fail, while GA wins over

GA is not a method to be applied for above circumstances for
comparisons of failure and winning.

It is suggested as alternate method.
Read complexity of solution space.

 

[curious_mind]

Bit of googling and came with three types of problems that can exist (and its mixtures). They are decision problems, search and optimization problems. There are tons of methods to address those issues. Being a novice, I would like to systematically elevate myself in these concepts. Could you suggest a reference, where traditional derivative or even non derivative methods are used to solve the problems and some comparison are made to provide merits and demerits of the methods used. I guess convergence would be the benchmark in comparison

 

[srizbf]

convergence is the criteria. But at what rate?
converging to acceptable solution (error delta) in
reasonable time.
That time is accepted based on the problem at hand.

 

[curious_mind]

Any good books with solved problems?

 

[curious_mind]

Adding to my previous question, how to choose the soft computing methods for a given problem?