Design, Establishment And Preliminary Application Of Digevol Digital Biological Platform In Biological Evolution Simulation Experiment

Digital organism is an important branch in the field of artificial life, which regards the programme as organism to conduct a series of biological characteristics activities, such as self-survival, self-reproduction, mutation, death, and competition with others. At present, the digital organism has become the most attractive research focus in the field of artificial life, has been extensively applied in the biological study including the life origin and evolution, genome variation, selection under environmental pressure and so on.In this study, we studied all the reported digital organism systems further and systemicly and found that all of them failed to achieve the free competition within the individuals and the natural selection which are critical for the difference of genome and phenotype in the process of biological evolution, although these systems have successfully simulated some basic biological phenomena such as survival, reproduce, mutation and death. The existing digital systems mainly depend on man-made rules and appointed scoring for the operating results to simulate the natural selection. The shortage of design conception makes the results of digital organism influenced by subjective awareness, and severely restrictes the application of digital biological systems, which turn into a major bottleneck for the development of digital organism.To date, there has not yet been the theory to explicitly describe the design and the requirement of the digital biological systems. In this study, we summed up the theoretical foundation of digital organisms and summarized the characteristics of self-organization system similar to the real life form, which should be an open and sustainable development architecture, and on this basis, we proposed six life traits required by the digital biological system, the digital organisms should be self-starting and self-survival, reproduction, variation, have life-span and withdraw all the occupied hardware resources after "death", individuals have a relatively closed but an independent architecture, and the hardware resources which are required to run program as a direct source of pressure from environment. We believe that these six characteristics are necessary for the life forms, the lack of any one can not become a sustainable development, an open self-organizing bio-mimetic system, and the reasons are interpreted in detail. To this theory, we design and build a new digital organism systems and named Digevol, meaning that Digevol is based on the sustainable evolution capacity of the order information (digital evolution).In the new system, we realized free competition and selection process within the digital individuals in the absence of the man-made intervention, which resolve the difficult problems successfully in the digital biological area. In the experiment, we made the digital organisms programmes directly compete the computer hardware resources with each other, by using the limit hardware resources as the environmental pressure, the digital programmes formed a spontaneous selection process and filter out some biological variations which adapt to the environment mostly. The competition and selection process in our design conception reduces the man-made subjective interference, and can simulate the real selection process in the natural environment more objectively. Meanwhile, because of getting rid of the man-made rules and appointed scoring for the operating results in Digevol, Digevol is an open system and the individual program is completely independent. Digevol increases many new degrees of freedom compared with the past digital systems, and also achieves many breakthroughs in bionic applications. In this study, we designed two different experiments to simulate evolution, one experiment studied the influence on the evolution direction, if the individuals had the difference in the competitive advantage during the process of biological variability; the other experiment discussed that the periodic changes in the environment had the select role for the hereditable individual activities cycle. In other experiments, we also try to simulate the select process under a variety of different environmental pressure, and construct the feasible test of multiple environmental niches.In the experimental to study the evolutionary direction of biological variability, we set up difficult variable capacity for the digital organisms to make sure that they can give birth to generation with different competitive advantages in the future. The evolutionary processes were divided into three categories, namely a continuing produce advantaged mutation, the phenotype with neutral mutation and produce the adverse variation. During the study, we compared each of the characteristics and laws in the three evolutionary processes, and examined the mutation rate, the difference of competition before and after variation, the number of intermediate type during evolution, and other factors such as group size impact on the evolutionary process. From our experiments, we can observe that, according to Darwinian evolution, the process of continuous accumulation of favorable variations is an irreversible evolutionary process with clear evolutionary direction, and the evolutionary direction of individual is always toward the mutation with more favorable competitive advantage, the speed of evolution is affected by the mutation rate and the difference of competitive advantage degree. However, in accordance with the neutral theory, the evolutionary direction is random and unpredictable, and the population size can change the distribution of different variant types. Finally, we observed that when adverse variation appeared in the evolutionary process, which would suppress evolution, but in some special cases, it is possible to reduce the bad effect result from adverse mutations.In the experiment to discuss that the environmental periodic changes had select role for the hereditable individual activities cycle, we set that the digital organisms can follow some periodic hereditable activities after birth, and the period length can be regulated through mutation and the mutation can pass down to the next generation. Our results indicate that environmental periodic changes had decisive role in the choice of individual activities cycle, and the strength of selection depend on individual its own activities amount within each cycle; intensity changes in the environment cycle; the duration of window supplying the hardware resource in environmental cycle.In this thesis, the Digevol preliminary application in the experiment proved that our design conception about the biological organisms systems is feasible, our systems successfully resolved the important defects in the existing systems which rely on the man-made rules and the appointed man-made scoring system to simulate the individual competition and selection process. The breakthroughs make the development of digital biological systems have a new starting point, and make it better serve biological research in future. Since the other bionic model and the mathematical modeling in the artificial life field also need evaluate fitness to achieve individual competition process and the environmental selection, which would not happen in the Digevol digital system. Therefore, our study is a successful attempt in the bionic conception.