Dear Emad Nabil. Thank you for expressing your point of view, complementing the question with some suitable examples of birds and insects (Cuckoo search, firefly and PSO).

Dear Antonio, I suggest you to avoid the metaphors and concentrate on the basics of the techniques.

Perhaps you may find the following article interesting

http://onlinelibrary.wiley.com/doi/10.1111/itor.12001/abstract

I agree with David. That is exactly the main reason for the question. I thank David also the link. I do not agree with Kenneth Sörensen classification of the algorithms. Some techniques as simulated annealing or GA are scientifically supported and do not deserve the classification of heuristic or metaheuristic (they are neither Deterministic nor Heuristic, nor Meta). Dear Emad thank you for the pdf.I thank both contributions.

Antonio:

   ga's and simulated annealing have some demonstrations about convergence that hold in very specific situations. I feel they are heuristic methods as they do not guarantee you to obtain the optimum in every possible situation.

Nevertheless this discussion is out of the original question.

David

Yes David. You are right in all points, except one. They "have some demonstration"; they "do not guaranty" the optimal solution in an isolated program run (they are not deterministic), and this "is out of the original question". They can be regarded as stochastic (or probabilistic) in the sense that they are able to get with high probability the absolute optimal solution after a random series of runs, with a relative low number of evaluations (when compared with the search space). If you are interested in this, please visit my RG page below.

Many people call the ancient Native Americans as Indians. Do you know why? How do you call a Native from India?

Article Optimal Layout of Sewer Systems: A Deterministic versus a St...

Non-Free Lunch Theorem tells us no algorithm is the best. we cannot design an algorithm that accommodates all problems best. Thus, we cannot conclude that one is more powerful. My viewpoint is that the answer do depend on the problem you'd like to address.

Take a look at 'Swarm Intelligence' book by kennedy&eberhart.

It compares genetic algorithms vs. social behaviors. From a fundamental viewpoint, it says genetic science could not describe evolution properly. But social metaphor gives a more extended vision of evolution through the definition of interaction between agents.

I think social behavior is more powerful because of self-organized management phenomena. self-organized systems have the ability to adapt themselves to the changing environment. 

Thanks Chen for bringing for this discussion that theorem. Thanks Bahram for sharing with us your opinion and the reference. However, I have to say that theory of Darwin sounds and has scientific support.

SI says that Darwin theory suffers from some drawbacks. But it doesn't deny Darwin's theory about evolution. Learning, Self-Orgainization and etc... are such points of controversy.

Thanks Bahram. According to your answer, is the opinion expressed on that book. There are many works on Metaheuristic methods. Here, I am more interested on the opinions and expertise of the RG members.

May be genetics approach yield few additions in terms of better characteristics in offspring. 'Social behavior' trend may be more useful for the current context in optimization & decision making. Most of the problem need fast solution  & algorithms based on uncertainties.

Res Sir,

Genetics may be the one, reliable in optimization context focusing over 'quantitative analysis & optimization'. And Social Behavior may be useful tool to predict 'qualitative parameters' in optimization problems to certain extent. I used GA tool for multiobjective optimization of Mechanical components.

Natural facts and phenomena always happens in an optimal way. That is why probably we try to mimic those to solve our problems by mathematical modelling. Those modelling are definitely not be consistent  since we don't know exactly what are process that actually happen in the nature. However, such inspiration can help us to design intelligent algorithms. Off-course such effort has resulted most successful and popular algorithms such as GA and PSO.  

Dear Anupam: Thank you for sharing your opinion. In fact many natural phenomena tend frequently for an equilibrium state, and the main difficulty is frequently the adequate modelling of the problems and the natural processes.

The most comprehensive comparison between genetic science and social metaphor has been discussed in (Swarm Intelligence book by kennedy and eberhart) as I know.

Genetics can help us understand why identical circumstances can elicit different behavioral responses among individuals. Genetic differences are reflected in variations in behavior; activation of distinct versions of a hormone receptor gene, an example Donaldson and Young present, results in monogamous behavior in one species of vole but not in another. Conversely, as Robinson et al. describe, insights from recent work show that perceiving social information—such as bird songs or dominance behavior from cichlid fish—from another individual of the same species can itself alter gene expression in the brain, with downstream effects on physiology and behavior.

http://www.sciencemag.org/content/322/5903/891

No empirical evidence is available to me, dear Antonio. I, therefore, cannot judge!

I think that there is no contradiction between genetics and social behavior; actually, they induce each other.

In some cases it contradict each other, but in most cases it is not. Rarely found cats live friendly with dogs and mice with cats. So social behavior induced by genetics.

Depending on the case, in terms of reliability and accuracy in finding the global optima, both methods are good. Upon my experience, the genetic algorithms are more general and powerful in formulating complex, multimodal and highly constrained engineering problems. However, the GAs are kind of slow and time consuming. Besides, the metaheuristics inspired by social behavior are computationally faster than GA. Honestly, I do not see other significant differences between these two methods.

Hi António Manuel Abreu Freire Diogo; I  agree to @ Prof. Kamal and @ Abedallah M Rababah  as DNA controls ones behaviour. Therefore genetics and social behavior both complement each other. Thanks

When it comes to behavior, we have moved beyond genetic determinism. Our genes do not lock us into certain ways of acting; rather, genetic influences are complicated and mutable and are only one of many factors affecting behavior.

Genetics can help us understand why identical circumstances can elicit different behavioral responses among individuals. Genetic differences are reflected in variations in behavior; activation of distinct versions of a hormone receptor gene, an example Donaldson and Young present, results in monogamous behavior in one species of vole but not in another. Conversely, as Robinson et al. describe, insights from recent work show that perceiving social information—such as bird songs or dominance behavior from cichlid fish—from another individual of the same species can itself alter gene expression in the brain, with downstream effects on physiology and behavior.

http://www.sciencemag.org/content/322/5903/891

Dear António,

Recent studies of specific genes supports the important influences of Genetic in behavior.

Heritability estimates suggest as much as one-half of the variation in propensity toward behavior can be explained by genetic differences among individuals. Moreover, various traits correlated with behavior appear heritable, including impulsivity, autonomic arousal, attention and cognitive deficits. On the other hand, Genetic research also makes a strong case for the importance of environment in influencing behavior. The environment itself appears to play a critical role in the expression and magnitude of genetic influences in behavior.

Research below, is about Genetics and it relation with the Antisocial Behavior.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2174903/

We've successfully used a fuzzy based architecture but have the rules and the membership functions determined using genetic algorithms. This approach has led to optimization in very large solution spaces. After learning is completed the resulting decision making is deterministic and not emergent. Therefore the outcome is verifiable and predictable. I feel that this is important for certain large scale, complex, safety driven systems.    

http://papers.sae.org/2014-01-2174/

Only in few cases it contradict each other, but in most cases it is not.

Thank you dear Kelly for your answer and the reference for your paper: “Learning of Intelligent Controllers for Autonomous Unmanned Combat Aerial Vehicles by Genetic Cascading Fuzzy Methods”. Fortunately you are speaking about “Unmanned Vehicles.” I do not know what your enemies of your paper are, but in my opinion the wars in general do not look very intelligent.  

Human intellectual and cultural evolution (social behavior) which are instinctive (e.g., mother-love, compassion, cooperation, curiosity, inventiveness and competitiveness ) took place much faster than genetic or physical evolution.  They were all necessary for the survival of the human.

From optimization point of view, an example of former is ICA (Imperialist Competitive Algorithm) and the later is GA (Genetic Algorithm). ICA which is an optimization algorithm based on a socio-political processes (i.e., based on human social behavior) seems to be superior than human's biological behavior (GA).

Dear Mahmoud. Thank you for sharing your opinion in this discussion. I agree with the first paragraph concerning to the natural velocity of the processes at a natural time scale. Although subject to some punctual historical reverse, freedom, solidarity, tolerance and plurality of opinions are extraordinary victories of the "human intellectual and cultural evolution".

Good contributions under different perspectives and levels of abstraction in this discussion. Thank you Dear Friends Mahfuz, Sarwan, and Jeanan by the good references provided. Dear Marwan, excellent criticism! Dear Kamal and Abedallah, wise equilibrium! Dear Hashem, who else! Dear Haghighi, intelligent answer!

Thank you all again

Dear Antonio, You were asking a question about "Several current optimization techniques..." and got good contributions already. I honestly could relate the answers and your comments to your question. Good Luck.

Thanks for keeping the discussion alive, dear Mahmoud. You can sincerely always count with my friendship for discussing ideas. The word empirical can perhaps have different meanings, since the context of the method until the experimental observation. For example, I do not call empirical data for experimental data or results of computing that were obtained through a scientific method. But the expression “empirical evidence” used by Marwan, for example, can perhaps be extended (validated) afterwards in a scientific approach (see the VIKIPEDIA link below – “Empirical Evidence”).

http://en.wikipedia.org/wiki/Empirical_evidence

Genetic behavior is original and influences and directed the social behavior. But, social behavior is  modified/adjusted behavior as per prevailing conditions and environmental stimuli.

The related RG discussion below about heuristic optimization methods might be interesting

https://www.researchgate.net/post/Which_heuristic_optimization_method_is_the_simplest_and_easiest_to_use

Can something/someone be truly optimized (maximizing the difference between benefits and costs) in a permanently changing environment?

Benefits can become costs (or vice versa) at different scales of analysis, e.g. population versus individual level or environment A versus environment B or...?

Behavior is dynamic -- Change even with generation change, may be little extent

Genetic behavior is having some permanency, but  Social Behavior adjust with change in social setting and environment

In any society, social behavior is more acceptable.

I do not think that the genes play an important role in the optimization context. It is the individual & social behaviors that direct the efforts towards the best conditions that  can be. Optimization is a science that ought to be learnt in every specialization & research ought to include more experimentation to seek optimal points to ensure continuous sustainable development.

Evolutionary biologists might claim that social behavior has a genetic basis, e.g. mechanistically guided by individual-environment interactions. Studies show for instance that animal personality profiles have a genetic basis also having possible consequences for social interactions and social organization (all this has been discussed before).

Why do species differ in social organization? Only genetically-based differences between species can explain this?

Dear Marcel: Thanks for your motivating challenges. In some areas and in some engineering problems, optimization is (or can be) an exact area of knowledge.

Dear Eraldo: My conviction is that we all learn always with each other (even if we sometimes do not notice it directly). I learned a lot teaching, for example.

Dear Antonio,

Perhaps aspects related to engineering are less dynamic, and therefore can be handled better from an optimization point of view?

In case of optimization in utilities, control on social behaviour is good. However in case of productivity increase, genetical approach is the right option. 

To answe this question is depending on the goal you need to achieve. The input variables therfore depends on the optimum contributionsfrom both with different percentage.

Thanks Dear Marcel for this pleasant small-challenge. Everything on earth is dynamic. In the optimization of an engineering problem I used Dynamic Programming and in the simulation of unsteady storm-water flows, I used a complete dynamic model, for example. The type of problem, and also/principally its correct modeling and formulation, and the selection of the suitable optimization technique, determines if you can solve it exactly by optimization or not (it does not need to be static in time or in space).

My thoughts are dominated at the moment by the problem of epigenetics, the activation and deactivation depending on the demands. This switching on or off a positions of the genom shows the influence of experience. So I have difficulties to differentiate which is dominant, genom or social learning. They are completely mixed.

Social behavior is another example of phenotypic plasticity and phenotypic plasticity is assumed to have a genetic basis.

Epigenetics also involves acquired characteristics in parents, e.g. through experience, and then transmitted to offspring (cf. Lamarck principles)?

Dear António,

Dear All,

Life situations are generally too complex and too complicated even to describe and to predict them. Optimisation processes use generally simplified models hoping to approach real situations. Unfortunately, neither genetics nor behaviour are sufficiently known. Complex behavioural patterns are often difficult to analyse. Most of participants are not specialists of the field thus contributions are rough approximations or guesses. Please mention some practical examples when contributing.

Example of complexity in simple social behavior:

The interpretation of wildlife behavior from a distance is not easy. Scientists cannot ask wildlife to explain in human language why they behaved as they behaved. Motivation expressed in the mind will influence behavioral action, but the same perceived behavior might result from different underlying motivations or feelings. This can be illustrated with scientific observations from a distance that one individual only drinks wine whereas another individual only drinks water. How to interpret this observation? Specialists in taste science might propose that the two individuals drink what they like to drink: one person likes wine, whereas the other person likes water or does not like wine. Specialists in social sciences may argue that the two individuals like wine, but that one of the two drinks water because the other prefers wine. Health specialists could perhaps propose that the individual that drinks water only drinks water because others told him that too much wine penalizes personal health condition. Wildlife specialists might explain that the person only drinking water does not want to share the bottle of water with the other person. Specialists in religion might think that the person that drinks wine is forcing the companion to drink water so that the companion remains healthy. Specialists in natural history might want to know the role of historical facts. Did the person that drink water drunk too much wine the day before? How can observers analyze and interpret wildlife behavior from a distance without having access to information provided by individuals explaining why they shared or not shared resources or without having access to historical facts prior to the initiation of observations? Apparently, there are many possible interpretations based on a few simple observations in the field or a cafeteria. Ideally, different hypotheses should propose different predictions perhaps requiring different tools or methods to test them. For instance, consumption constraints could be decreased by giving two individuals >2 bottles of wine and >2 bottles of water. The order of consumption of all the bottles received might test the ‘taste’ hypothesis in more detail or could verify if two individuals exchange bottles differing in content, and if so, in what direction. Blood tests or other physiological sampling methods might indicate overconsumption of resources (e.g. wine) in the recent past or in the absence of observers. Wildlife psychologists or ethnologists might accept or reject some hypotheses through analysis of face expressions, gestures or social interactions.

And then we did even think about genetics in social interactions

Dear Marcel,

I hope also the other participants enjoyed like me your pseudo-Cartesian comment.

Dear Andras,

I realize the example is a bit long, but, on the other hand, other explanations from other disciplines could have been added to make the list of explanations concerning a simple social interaction a bit longer.

Cheers

Solution analyzes the social web and raises events related to intelligent owned and earned channels, conversations, arguments and audience monitoring every relevant content and interaction for the brand. The platform processes millions of web and social data on which you can build reports and alerts based on geographical, demographic, psychographic and relevance in order to optimize the planning and operation of their digital marketing.

http://social-intelligence.bewe.co.uk/

Dear Hashem: I suppose I understood your answer and I agree. For example in a GA the mutation rate, that reflects the adaptation of the species to the environment, can be higher or lesser depending of the dynamic of the environment and of the possible rapid loss of genetic material that can be important for the fitness of the populations in further generations.

Although the deficit in mating between mice of the same MUP type was very strong, there was no evidence that animals used genetic markers other than MUP to improve their level of inbreeding avoidance. Further, there was no evidence to support the previously untested hypothesis that animals could increase the range of relatives avoided by imprinting on the separate haplotypes carried by their mother, which also results in the rejection of any unrelated animals that happen to share one maternal haplotype.

https://www.google.co.in/url?sa=t&rct=j&q=&esrc=s&source=web&cd=61&cad=rja&uact=8&ved=0CBwQFjAAODw&url=http%3A%2F%2Fwww.sciencedirect.com%2Fscience%2Farticle%2Fpii%2FS0960982207021380&ei=y1UJVYifOtGWuATI64DwBA&usg=AFQjCNFV7Qs7_fbDb17FoyEpO7gO2j4syA

Thanks Dear Krishnan for bringing us a little excerpt and the link to the paper from 2007 by Amy L. Sherborne et al “The Genetic Basis of Inbreeding Avoidance in House Mice”, Current Biology, 17(23), 2061–2066.

Dear Marcel: Thanks for your good (humorous) example of complexity of social behavior regarded by several different perspectives. By the way, I suppose that the wine preference also depends of the circumstances, meal and quality of the wine (with a good Portuguese wine in a good meal, for example, rare people would prefer water!). And if you add other drinks, and mixtures, the complexity of your simple example would increase substantially.

Dear Andràs: Thanks for your answers. I agree when you wrote: “Life situations are generally too complex and too complicated even to describe and to predict them. Optimization processes use generally simplified models hoping to approach real situations.” However, Biology and Human Psychology are certainly complex fields a little out of my scientific reaching.

In my opinion, the art of engineering is to establish good approximation by considering the more relevant aspects and ignoring the less relevant aspects (of course when it is not possible considerer all aspects). Let me give a very simple example: Between any two numbers real (even if very, very close), there is an infinity of numbers real. For constructing in practice something previously projected, is always necessary some sort of approximation (even if preferentially minimal). But this does not harms in any aspect the validity and usefulness of any optimization procedure.

Genetics

In optimization context, both genetics and social behavior can be powerful and reliable. Genetic algorithms are a type of optimization algorithm that is inspired by natural selection. They can be used to solve complex optimization problems that are difficult to solve using traditional methods.