Research And Application Of Intelligent Optimization Algorithm Based On Foraging Behavior

The diversity, complexity and intelligence of biology for modern industrial technology innovation provide unlimited space bound to advance the biological and intelligent industrial technology revolution. Foraging behavior is the most basic behavior in nature and it is necessary for the survival and reproduction. This research is based on the three aspects:the individual adaptive foraging, the information exchange and the life cycle search, the multi-group collaboration. We attempt to create the simulation of the inherent law and evolution mechanism of the nature, and then to construct a unified framework model of the biological foraging optimization algorithm. In this paper, many research results have been made covering the theoretical analysis and engineering application in the research of bio-inspired, which will be described as follows:(1) Based on the theory of the concentration of plant morpheme and L-system, we simulate the growing status of genuine plant root systems, analyze the growing and foraging behavior of genuine plant root systems through simulating, and finally propose an adaptive plant growing optimization model and approach-Root System Growth Optimization (RSGA). Sphere and Griewank functions are applied in testing the hydrotropism and gravitropism of root system. The simulation results testing against the set of CEC2005 test function show that the proposed RSGA mode has a good performance on convergence accuracy and convergence speed. Especially for high-dimensional problems, RSGA has a high efficiency. For solving the practical engineering applications, such as continuous and dynamical problems, this proposed algorithm is a new method.(2) On account of the research about information communication mode and life period searching approach, we simulate the clone, split, and death of bacteria, systematically study the simulation modeling of the behavior of typical bacteria group, and propose a novel optimization algorithm of bacteria group foraging. So we propose a new algorithm:Life-cycle Bacterial Colony Foraging Optimization, LBCFO. The tendency curves of population size changing against Sphere, Rosebrock, Rastrigrin and Griewank test functions show that the "first largening, then diminishing" meets the law of microbial life change phenomenon. For testing the basic benchmark functions, the results show that LBCFO has a better convergence rate and accuracy than othtr BFO's derivatives. The mechanism of lifecycle improves its adaptability and efficiency. In order to verify the performance of the proposed new type of bacterial swarm foraging algorithm for complex engineering optimization problems, the experiments of 3D printing system in intelligent manufacturing field have been tested. The mode of printhead is structured in ANSYS software. LBCFO got a smaller error towards target droplet velocity and volume than BFO, BSO and ABFO. The experimental results verify the feasibility and effectiveness of the proposed algorithm.(3) Based on the simulation of the layer topology structure of complex adaptive biological system in nature, and the communication rules among distinct individuals, groups, and layers, we make up a layer multiple honeycomb optimization algorithms according to the relationships between single group and multiple groups Multi-colony Coorperation Bee Foraging Algorithm, MCBFA. Through adopting some classical topologies into MCBFO algorithms, the directions and speed of information exchanging is controlled. For testing against Rosenbrock, Ackley, Rastrigrin and Griewank test functions, the results show that coevolution strategy can improve the diversity of population. In orde to test its ability of solving practical engineering problems, the real multi-threshold image segmentation is applied. After testing the basic image segment set, the results show the proposed coevolution algorithm can effectively overcome the backdraw of the premature convergence problem of the traditional single layer biological heuristic optimization model, and further improve the convergence speed and convergence precision of the intelligent algorithm.(4) Analogying the self-adaptive foraging phenomence in nature and dynamically route planning of mobile robot, we design a new biological inspired computing algorithm-Dynamic Animal Foraging Optimization, DAFO. This algorithm introduces some natural biological local search strategies and adaptive foraging strategies. For testing against some unconstrained complex dynamic multi-modal test functions, the results confirm that the proposed DAFO algorithm has high accuracy and stability, and has the capability of dynamic optimization. Using the Sphere function as the simulation environment for Robot route planning optimization, searching subject driven by DAFO can smoothly avoid obstacles, quickly find the target site and effectively saves walking time. It can be concluded that, our bionic intelligent optimization approach is effective, stable, and competitive, and it performs excellent as to the searching efficiency and solution accuracy in the solving of complex engineering optimization.