Research On Biological Inspired Intelligent Computation And Its Applications

Many kinds of complicated problems which exist in scientific calculation and engineering calculation can be summarized as optimization calculations essentially. One of the bases of support vector machine which can be used for classification and regression is optimization calculation. A large number of objective functions exist in the optimization problems. The differentiable functions are always continuous while the continuous ones are not always differentiable. A lot of discontinuous non-smooth functions exist in nature objectively. Even though it is a continuous function, it may have a lot of differentiable points. Some objective functions are always continuous here and there but can not conduct the derivation. The traditional optimization algorithms include the filled function algorithm and tunnel algorithm. For a large number of objective existing non-smooth and discontinuous functions, they have difficulty in making a satisfactory optimization effect in limited time.With the deepening of the modern scientific technology research, the researchers find and put forward a lot of optimization problems of high dimension, nonlinearity and nondifferentiability. How to solve the optimization problems, especially the problem of combinatorial optimization, is the question which has been troubled people all the time. Traditional optimization theories based on the functional derivation, obviously, can not meet the demand of the optimization problems. In order to solve the optimization problems which can not be solved by the traditional algorithms, researchers put forward some random optimization algorithms and one of them is biological inspired algorithm. These biological inspired algorithms have gotten a good effect in many problems. Therefore, this paper works to study biological inspired intelligent optimization algorithms.The biological inspired intelligent optimization algorithm is a kind of intelligent algorithm put forward because of the enlightenment of some behaviors of creatures. We conduct the research by the means of combining the theory innovation and simulation experiment. The biological inspired intelligent optimization algorithms discussed in this paper have difficulty in making a satisfactory optimization effect in limited time for some high-dimensional non-smooth and non-differentiable functions.In this paper, the characteristics and innovative points include:First, make a collocation and classification for the test functions. Second, improve the exiting intelligent biological inspired optimization algorithms. Third, apply the biological inspired algorithm to some fields and solve many problems, set up the three-dimensional coverage model of sensor nodes in Internet of Thing and put forward a three-dimensional deployment algorithm in Internet of Things enlightened by the crowd cooperation searching.In the nature, there are many kinds of creatures. Since they can survive after a long period of natural selection progress which selects the superior and eliminates the inferior, their existences are meaningful naturally. In the nature, human beings and many creatures are coexisting harmoniously and different creatures have different characteristics. Now that they can survive, the creatures have their specific characteristics. If our intelligent algorithm can absorb the merits of many kinds of creatures to serve people, it will become very meaningful. The main work contents of this paper include:First, we make a collocation and classification for the test functions. A unified standard is needed to compare the quality of the algorithm and program. With the standard, we can make a judgment about the superiors and inferiors of different algorithms. For the optimization algorithm, the standard which is used to test the superior and inferior of algorithm is the test functions. However, there are millions upon millions of target functions. In order to make the comparison conveniently, many scholars use a lot of benchmark functions and evaluate the superiors and inferiors of these algorithms on account of the testing results of the benchmark functions. According to the "NO Free Lunch" theory, for different test functions, the superior and inferior of algorithm is also different. Therefore, we can not use one or two functions to take place all optimization problems as a unified standard. The actual optimization has diversity, therefore, as the standard of testing optimization algorithm; the test functions also should have diversity. In different papers, the used test functions may also different, which bring inconvenience to the comparison. In order to facilitate the design and improvement of the optimization algorithm, we collect and refer to a lot of literature materials and make a collocation and classification for the test functions.Second, we improve a lot of intelligent biological inspired optimization algorithms. In this paper, we put forward a kind of improved particle swarm algorithm, which strengthen the memory and learning ability of the particle swarm. In order to accelerate the convergence, we introduce the mechanism of the survival of the fittest and conduct experiments under the same parameter condition. The experiment result shows that the global optimal solution found by the improved particle swarm algorithm put forward in this paper coincide with the theoretical optimal solution of the test function accurately. Comparing with the basic particle swarm algorithm and second order particle swarm algorithm, the algorithm put forward in this paper have good global optimization ability. According to the brains storm algorithm, we put forward the group brains storm algorithm based on orthogonal design. In this paper, we also conduct experiment by a lot of test functions. The experiment result shows that, for this test function, the orthogonal brains storm algorithm is superior to the original brains storm optimization algorithm.Third, we apply the biological inspired algorithm to some fields and solve many problems. In this paper, we put forward a biological inspired image segmentation algorithm with time control facing the embedded system. We also apply the improved particle swarm optimization algorithm to the image segmentation and give the related experimental results. We set up a three-dimensional coverage model of sensor nodes in Internet of Things and put forward a three-dimensional deployment algorithm enlightened by crowd cooperation search. We improve the Virtual Force Algorithm and expend it into the three-dimensional space. Enlightened by this kind of crowd cooperation search, we improve the gravity of sensor nodes on the target point and put forward the three-dimensional deployment algorithm based on the crowd cooperation search. Many traditional sensor network coverage models use the additive models based on the probability measure. Because of the influence of the environment change and many other factors, the detection radius of the practical sensor node is usually changing, In order to solve this problem, we introduces g-λ measure, 2additive measure and K additive measure. We modify the Lenard Jones potential and construct the simulation molecular force, we also introduce the variable and adjustable maximum displacement, boundary gravity and the adjustable maximum displacement concerning the coverage rate. Many kinds of biological inspired deployment algorithms based on evolutionary computation are put forward in this paper. Afterwards, we make a comparison between the improved deployment algorithms enlightened by crowd cooperation search with the biological inspired deployment algorithms and conduct comparative tests by making different awaiting detection areas. The experiment result shows that the deployment algorithm put forward in this paper has the best optimization effect, comparing with the deployment algorithms that are basing on particle swarm optimization, DE, brains storm, GA, improved random walk, simulated annealing and group simulated annealing. At last, we make a regression and fitting on the coverage rate gotten by this deployment algorithms. According to it, we carry on a forecast on the network coverage rate with unknown nodes quantity and get a satisfactory effect.In conclusion, we study biological inspired intelligent optimization algorithms in this paper, apply the biological inspired algorithm to the digital media field and solve many problems. Compared with other algorithms, the algorithm put forward in this paper can get a better result. The biological inspired intelligent optimization algorithm has an important role in a circuit system design, motor, wind tunnel of plane and so on. In all aspects of the social production and life, the biological inspired intelligent optimization algorithms will have a huge development space.