Multi-Objective Optimization And Evolutionary Algorithm`s:)

[englishrick]

The term evolutionary algorithm "EA" stands for a class of stochastic optimization methods that simulate the process of natural evolution....since the 1970s several exiting evolutionary methodologies have been proposed, mainly genetic algorithms, evolutionary programming, including evolution strategies......

hopefully in this thread,,,we will cover all aspects of multi-objective techniques,,including the math behind multi dimentional theory

 

[englishrick]

the main objective of my plant studies is to link chromosome loci to specific agricultural traits in the hope of increasing breeding efficiency....recently developed marker-assisted selection strategies rely on attempts to identify and quantify the genetic contributions to the phenotype...,,,,,

identifying quantitative traits is a tricky business ,,,,it requires special strategy "ricks special strategy" lol ,,,,[not realy],,,,,,,, the "overall" strategy is to firstly to develop a population ..this will be the primary source of tissue to isolate DNA or protein from.......later breeders [and rick] can increese numbers and start "mapping population",,,, segregating for the target traits and molecular markers begins..yeehaa .

there are many types of seedlines,,,"every seed is not made equal" lol,,,,,,,,,backcross populations are developed by crossing the F1 with one of the two parents used in the initial cross.... the major drawback with exclusivly growing F2 or backcross populations is that them populations are not eternal lines!!!.....your source of tissue to isolate DNA or protein will be exhausted at some point in time!!!!!!! ....you then would have to begin mapping again in another population.....imo that would just be vicious cycle!!,,,,an i hate vicious cycles!!.....populations of recombinant "InbredLines" can be a powerful solution to this problem..... recombinant "inbred lines"...or... "pure lines" are developed by single-seed selections from individual plants of an F2 population!!!...

 

[englishrick]

evoloution is acoplished by organisms rearranging genetic material to survive in environments that confront them...

in 1992 John H. Holland published: "Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control and Artificial Intelligence",,,,,,,,,,,,Holland is is the farther of genetic algorithms!!,,,,,the book he wrote in 1992 "Adaptation in Natural and Artificial Systems" is the book that initiated the "EA" field of study!!!


*EA = Evolutionary Algorithm

 

[englishrick]

holla if ya want more?

 

[VerdantGreen]

evoloution is acoplished by organisms rearranging genetic material to survive in environments that confront them...

....

not so sure about that rick. organisms dont rearrange their genetics, what happens is that all organisms are slightly different - even within a species - so those whose differences that give them an advantage within their environmental niche will be naturally selected and thus will be the ones that get to breed and pass their advantageous genes on to their offspring. the flipside of this is specialization. organisms that evolve to be very successful within a certain niche environment are up sh1t creek when that enbvironment changes or dissapears. think pine martins or wildcats in the UK.

so multi- objective optimization is like, making things the best thay can be from the widest array of perspectives ??

V.

 

[englishrick]

niche ,,,,

would you say polymorphism comes under "organisms rearranging genetic material to survive in environments"??

 

[Blimey]

evoloution is acoplished by organisms rearranging genetic material to survive in environments that confront them...

You make it sound as if organisms deliberately manipulate their genetic makeup. Not so.

Mutation of genetic structure occurs thru such events as

a) Chromosomal damage casued by the presence of a mutagen in the environment (e.g. naturally ocurring radioactive minerals, exposure to certain wavelengths of electromagnetic spectrum, colchicine)

b) Errors in the DNA copying process - it's almost perfect, but not quite

The majority of mutation is either non-beneficial (in which case it is unlikely to spread within the genepool), or at best makes no difference, in which case it will most likely remain fairly isolated.

Any mutation (or combination of mutations) that gives even a minimal advantage, will likely spread and become more common.

Combine that with isolation of a population containing those beneficial mutation(s), allow many many generations (i.e. lots of time for most multi-cellular organisms) and you get speciation.

 

[englishrick]

hay B

plants oviously dont seem to have mutch choice in there mating habits like we do,,,,,,,,BUT, like you said, plants are subject to enviroment,,,,changes in eviroment can instantly impact on expreshion!!,,,,,lets take Spontaneous Revercal for example, "things dont actualy change they just express a change", this can and will move things in a new directions untill the enviroment demands change again ,,,,,,,no?

 

[englishrick]

not so sure about that rick. organisms dont rearrange their genetics, what happens is that all organisms are slightly different - even within a species - so those whose differences that give them an advantage within their environmental niche will be naturally selected and thus will be the ones that get to breed and pass their advantageous genes on to their offspring. the flipside of this is specialization. organisms that evolve to be very successful within a certain niche environment are up sh1t creek when that enbvironment changes or dissapears. think pine martins or wildcats in the UK.

so multi- objective optimization is like, making things the best thay can be from the widest array of perspectives ??

V.

imo,,,the general idea of multi-objective optimization is to provide soloutions to niche enviroments

 

[Blimey]

plants oviously dont seem to have mutch choice in there mating habits like we do,,,,,,,,BUT, like you said, plants are subject to enviroment,,,,changes in eviroment can instantly impact on expreshion!!,,,,,lets take Spontaneous Revercal for example, "things dont actualy change they just express a change", this can and will move things in a new directions untill the enviroment demands change again ,,,,,,,no?

Yes...

Two mechanisms (modification of genotype, and expression of genotype subject to environment) both subject to Natural Selection.

In order to differentiate betwixt the two - we again enter the old familiar debate re breeder practices.

e.g. I (and most breeders (not that I consider myself one)) don't have the resources to take a gazillion clones of each plant and run them under every possible combination of lights, soil or soilless media, watering regimes, feeding regimes etc etc etc.

No-one has. One reason why the better breeders (no names, no packdrill) send out a shedload of seeds for us to play with prior to general release.

The same clone running in my small flourescent box using whatever nutes and soil media I can scavenge for the lowest outlay will NOT express identically as when you are running it in your hi-tech, HID NFT environment.

Even with lots of beta-testing, only the tip of the iceberg is scratched, and it's a costly procedure to have to go back to the drawing-board, as it were, if (for example), said clone routinely hermies like crazy in my environment, but produces ic420 cup-winning smoke 100% of the time in yours.

It's the same reason why that expensive piece of software you bought runs like perfect on your PC, but causes all hell to break loose on mine.

And people who have chosen to make their living as breeders...need to make a living. They need to move product, imperfect as it is.

In practise.....does it matter whether it's "bad genes" or "genes that express badly under condition X" ???

I dunno....but I'd like to hear other people's views.

 

[englishrick]

i think we have a good debate going here,,,,,,

i think i might be able to bring you to agree with me B,,,,,im gona try anyways,,,,,

B-----how do you feel about Evolutionary Algorithm`s ?,,,,i think its probbly an amzing tool and a sound concept, what do you think B?


i would also like to hear from more people.

 

[Blimey]

To be honest....it's not a term I'm familiar with.

Are we talking computer methodologies? I'd have to look it up and get back to you once I've had a chance to digest.

Here's a little something I was thinking about a few minutes ago while heading down the shop to get my baccy:

Given that we know that a certain set of genes - (ideally a clone, but we should be able to generalise to a certain degree about a fairly reliable seed variety) - will express optimally in certain environments and sub-optimally in others....

Why is pretty much the best clue to this that you will get from a seed bank is...."Indoor or Outdoor"? or "good for SOG"?

I've never seen a strain advertised saying...."developed for growing under Compact Flourescents...will not perform as well under HID"

or "performs best in coco using a high fulvic acid regime....use of (whatever nutrient) is not advised for optimal results".

Because they don't know - due to lack of beta-testing....or because it would hurt sales - regardless of impact this would have on customer satisfaction?

At the moment, the best we have is anecdotal evidence from The "Strains & Hybrids" forum, or the "Grow Diaries" and "Seed Vendors" fora.

And anecdotal evidence from a single grower, where so many variables are unknown is....not so hot. TBH, It's neither fair nor realistic to expect hobby growers to provide data of the depth and quality we need to make truly informed decisions, especially as many are so damn high while posting

Apologies if I'm wandering OT....no intention to hijack.

 

[Buffoonman]

Wow youve certainly picked a complex thing to debate. Nature has so many variables, are Evolutionary Algorithm`s a way of modeling evolution ? I like the way you were pondering this whilst walking to get some baccy.

 

[Blimey]

Darwin developed most of his ideas walking endless circuits around and around his (extensive) garden at Down House in Bromley (not at all far from me, actually) ....if it's good enough for him.... an idiot like me can always use inspiration from mental giants like Darwin.

 

[Buffoonman]

I wonder if weed will get stronger as its in its own interest for survival. i.e man will be more likely to cultivate it.

 

[Blimey]

I wonder if weed will get stronger as its in its own interest for survival. i.e man will be more likely to cultivate it.

Once man gets involved, Natural Selection goes out the window, and Artificial Selection is the process.

In Richard Dawkins most excellent "The Ancestor's Tale" (a great book, for those looking for a heavy-duty Xmas read), p.218 he looks at a study, started in 1896 at the Illinois State Agricultural Laboratory. They started with a strain of maize containing (average) 4% oil. They subsequently selected only on the basis of oil production for 90 generations (90 years), at the end, the (average) oil % was 19%.
Study published in 1992 by J.W Dudley and R.J Lambert for those with an interest.

I don't know what other traits in this line...such as vigour, were like after 90 generations.

My point is...man can do in a relatively short space of time, what Natural Selection will take tens of millenia to do, if it does it at all.

Sorry for heading OT there, Rick, but I thought it might help illuminate a point.

 

[Buffoonman]

Surley mans influence is a part of nature however strange.

 

[Blimey]

ahahahaha.....philosophical debate. I love it

You know.....I've been sitting here for a few minutes now, typing stuff in, thinking about it, rubbing it out...trying to give you the answer that will prove my point without getting all metaphysical, or indeed talking about N-dimensional hyperspheres and special relativity, as a certain acquaintance of ours has a tendancy to do.

But my muse is not here. No weed and no hashish make Homer go crazy. Let me get back to ya.....

 

[englishrick]

hahaha^^^lets get quantam

BOOF is providing good stimulation,,,,,,,thanks Boffman

i agree,,,,,,people create!!,,,,,people are an offshoot of nature,,,,,,



BLIMeY----its impossible to explain "removed enviroments" without quantam phisics,,,,,,,its all about hypervolume bro,,,,,,once you turn every dimention into math its only explainable with n-Space.....Euclidean type geometry,,,,,,the home of special relitivity

the reason why you dont see seedbanks saying "Made for LED`s" or "low imput enviroments" is because they are realy just hacks,,,,,,,,the seed biz comes in 2 levels amsterdam coffeshops and pharma companys,,,,,,comon ground is hard to come by blimey bro

the young stuff is F1 and Backcross,,,,,,,,but the source will exaust,,,,,,,,,, the big boys use MULTI-OBJECTIVE OPTIMIZATION with Evolutionary Algorithm`s,,,,the old boys want to keep traits forever, in all enviroments an make things self sustainable,,,,,,,,,

it seems to me,,,,,,,,,the only way to make traits breed true, is to make the respecting geinepairs homozygote within a chosen inderviduals,,,,,im proposing to do this via an Evolutionary Algorithm



im gona try to anyways

 

[Blimey]

Try away, Rick...I'm here to learn too.

But...baby steps, please.

It's when you put on those 7-league boots and a post seems totally random and unconnected in any way to any reality we may both inhabit (or indeed the topic of the thread) that people go "WTF?....Stoner" and just tune you out.