Читать книгу Evolutionary Models in Business - Evgeny Klochkov - Страница 6

Biological evolution (from Lat. Evolutio – “deployment”) is a natural process of development of living nature, accompanied by changes in the genetic composition of populations, adaptations, formation and extinction of species, transformation of ecosystems and the biosphere as a whole.

Charles Darwin was the first to formulate the theory of evolution by natural selection in his work On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life. He proved that variability and heredity are common properties of all organisms. Due to intensive reproduction, organisms create a sufficient amount of raw material for the selection of the “best” by destroying (eliminating) the “worst”. Moreover, natural selection functions in the presence of two factors – the intensity of reproduction and the struggle for existence. The struggle for existence inevitably follows from the high speed with which organisms increase their numbers. Further, Ch. Darwin suggests that since more individuals are produced than can survive, there must inevitably be a struggle for existence, either between individuals of the same species, or between individuals of different species, or with physical living conditions. Let us consider each type of struggle separately.

Interspecies struggle means a competition for survival between individuals of different species. It has a complex nature and manifests itself in the following types of harmful and beneficial relationships:

a) competition means any antagonistic relationship associated with the struggle for existence, for domination, for food, space, and other resources between organisms, species, or populations of species that need the same resources;

b) predation means a phenomenon, where one organism feeds on the organs and tissues of another, while no symbiotic relationship is observed. It is worth noting that killing the victim is optional;

c) parasitism means a form of symbiosis, where one organism (the parasite) uses another one (the host) as a source of food and/or habitat, while imposing (partially or completely) the regulation of its relations with the external environment upon the host. There are also an obligate form of parasitism, when the parasite cannot exist without the host (e.g., viruses are a typical example), and an optional form of parasitism (e.g., lice, fleas, parasitic worms, etc.);

d) commensalism means relations, where one species, without damage or benefit to itself, contributes to the prosperity of another species (for example, sheep and cattle spread plant seeds on their wool);

e) mutualism means relations, when two species mutually support each other (for example, insects and birds pollinate flowers; cereals and legumes contribute to the growth of each other in grass mixtures).

Intraspecific struggle includes the relationship between individuals of the same species with similar needs for food and territory. It is of the most acute character, since representatives of one species, especially one population, require the same conditions for life and reproduction of the offspring. For example, red cockroaches completely displace black ones, a gray rat displaces a black one, a European bee displaces an Australian one. Competitive intraspecific relationships are widely known everywhere. Birds of the same species compete over nesting sites. Males of many species of mammals and birds enter into a struggle with each other for the possession of a female during the mating season. An excessive increase in the population size exacerbates the struggle for food, therefore, for example, cannibalism, i.e. eating individuals of one’s own species, is widespread among fish. In the process of evolution, many animals have developed certain adaptations that help them avoid competition with other individuals of their species, e.g. marking the boundaries of their site, life in herds, and threatening postures.

Fight against unfavorable environmental conditions. A huge number of plants are destroyed almost every year by late frosts, droughts, and sharp climatic fluctuations. The mass of seeds is carried by the wind into unfavorable conditions and perishes. Many animals die during severe winters with little snow. With a lack of oxygen in the water, fish are killed. The result of this struggle is the survival of individuals with the most favorable hereditary changes for the given living conditions. For example, desert plants have long roots and small leaves.

All of the above types of struggle for existence lead to the extermination of a huge number of individuals or to the impossibility of leaving offspring. As a result of natural selection, individuals who are most adapted to the environmental conditions in which this species lives, survive. At the same time, there are no organic forms absolutely perfectly adapted to the conditions of their life in nature. This is impossible given the variability of the environment. Having become of no use, organs that were originally formed under the influence of natural selection, can show great variability under the influence of new environmental factors. When the conditions that formed this sign alter, it may turn out that what was useful will become harmful. Therefore, the idea of relative expediency in organic nature arises. The evolutionary process is always adaptive. The polar bear can serve as an example of survival by natural selection. Under the conditions of low temperatures in the Far North, he has formed a special wool with hollow hairs that have high thermal insulation properties. The soles of their feet are lined with wool to prevent slipping on the ice and freezing. There is a swimming membrane between the toes, and the front of the paws is trimmed with stiff bristles. Big claws can hold even really strong prey. This is a necessary and sufficient condition for survival. Individuals with other mutations simply did not survive. They could not stand the harsh climatic conditions, competition with each other, and competition with other species.

However, in conditions of continuously fluctuating environmental parameters, adaptive mechanisms are not enough. Nature has provided for another way to adapt to external conditions, namely, generational change. This mechanism guarantees the survival of the species. An organism must be born, master the living space, bring in something new that is conducive to survival in the development of the living space, reproduce offspring, transfer this new acquisition to them, help them get used to the living space and die itself. Thus, the key idea of survival is that organisms must be in a state of constant adaptation to the pulsating external conditions, acquiring some mutations and passing them on by inheritance, thereby fixing effective survival mechanisms in the genetic code.

In the process of research, Ch. Darwin noticed and made another interesting practical conclusion, which is also applicable for the world of business: natural selection can be supplemented by artificial selection. In his work, he described the process of creating new breeds and varieties of cultivated plants with properties and traits over a number of generations tha are valuable to humans. As a result, artificial selection has received an important practical application, when breeders set a task and carry out selection according to several criteria, breeding plant cultures and animal breeds with given properties and characteristics. As one of the examples, Ch. Darwin speaks of farmers in Virginia whose pigs were all black. When asked why, they informed him that the pigs were eating dye roots (Lachnanthes), which caused their bones to turn pink and all but the black varieties lost their hooves. And one of the farmers added: “In each litter we select black piglets for raising, as only they have the undoubted opportunity to survive.” Gradually, this theory developed to stimulate further dynamic development of the science of selection, the theory of mutations, the theory of gene structure, and the molecular basis of heredity. Breeders obtain the necessary properties through targeted selection of starting material for breeding, hybridization, mass and individual artificial selection. For instance, when breeding animals, breeds with high productivity, vitality, resistance to diseases, and adverse environmental conditions are developed.

An amazing and expensive cat breed called the Savannah can be mentioned as an example of targeted breeding. Being a home variant of the wild serval, the savannah was bred in the 1980s. Wild cats have always been popular with the elite, and in order to protect the true cheetahs and leopards, breeders have created an alternative. The animal looks formidable and dangerous, but in fact it is affectionate and sociable. The first savannah was introduced to the world in 1986 by the Bengal breeder J. Frank. It was the result of crossing a true serval male with a domestic Siamese cat. And in 2001, the breed was officially recognized and registered.

Darwin’s theory has been criticized many times, but the idea that life developed rather than was created in a “ready-made” form does not raise doubts among the overwhelming number of scientists. One of the practical conclusions from the theory of evolution was the emergence of genetic algorithms proposed by

J. Holland in 1975. Genetic algorithms will be very useful to us for using within the framework of the paradigm proposed by the author, and their application will be discussed below.

Genetic algorithms are adaptive search-based techniques that are used to solve optimization problems. They make use of both an analogue of genetic inheritance mechanism and an analogue of natural selection, at the same time preserving biological terminology in a simplified form and basic concepts of linear algebra. The first genetic algorithm scheme was proposed by J. Holland at the University of Michigan in 1975. And Ch. Darwin’s Theory of Evolution along with the studies of L. J. Fogel, A. J. Owens, and M. J. Walsh on the evolution of simple automata intended to predict symbols in digital sequences served as the premises thereto. The new algorithm was named Holland’s reproductive plan and was later actively used as a basic algorithm in evolutionary computations. Holland’s ideas were developed by his students K. De Jong from George Mason University in Virginia and D. Goldberg from the Illinois Genetic Algorithms Laboratory. Due to them, a classical genetic algorithm was created, all operators were described and the behavior of the test functions group was investigated. It was Goldberg’s algorithm that was called the “genetic algorithm”. To understand the essence of genetic algorithms and the opportunities of their application in business, it is worth dwelling in a little more detail on the stages of this model.

Phase 1. Creation of a new population.

In the first phase, an initial population is created. Requirements for the quality of the population according to the given parameters are not critical, since in the end the algorithm will correct this problem. The most important thing is that the population should conform to the “format” and be fit for reproduction.

Phase 2. Reproduction.

It is important that the descendant (child) should be able to inherit the parents’ traits. In this case, all members of the population reproduce, and not only the survivors. Otherwise, one alpha male will stand out, whose genes will supersede all the others, and this is fundamentally unacceptable.

Phase 3. Mutations.

Mutations are similar to reproduction. A certain number of individuals are selected out of the mutants and modified in accordance with predetermined operations.

Phase 4. Selection.

At this stage, the most important process starts. The experimenter selects from the population the proportion of those who will “go further.” The proportion of those who “survived” the selection is determined in advance as a preset parameter. Then the cycle is repeated from the beginning. If the result is not satisfactory, these steps are repeated until the result becomes satisfying or until one of the following conditions is fulfilled: either the number of generations (cycles) reaches a pre-selected maximum, or the time allotted for mutation gets exhausted.

Genetic algorithms are referred to as soft computing. The term “soft computing” was introduced by L. Zadeh in 1994. This concept brings together such areas as fuzzy logic, neural networks, probabilistic reasoning, trust networks, and evolutionary algorithms, which complement each other and are used in various combinations or independently to create hybrid intelligent systems.

To put it simply, a genetic algorithm is a method of item-by-item examination within the available set for selection of only those solutions that meet the given parameters. The solution of the task to find the shortest path from point A to point B can serve as an example here. First, you need to build any possible paths from point A to point B, look for all possible options, shift existing route points, add and remove points, through which the route passes, i.e. create the maximum number of options. Then, to choose the shortest path, you need to calculate the lengths of the routes and compare them. In the process of searching, we discard some of the longest routes, compare the remaining ones, and in the end a single route, i.e. the shortest one, is left. Only practice can show which approach will be correct in the context of a specific task.