Research Swaron Quantum M Optimization Algorithm

When solving the problem of various variables in theory and practical application,many intelligent optimization algorithms have defects of slowly convergence and easilygetting into local optimal value. Artificial bee colony algorithm has the advantages ofless parameters, simple calculation, fast convergence and strong robustness. Thealgorithm has been applied into function optimization、 combinatorial optimization andengineering field widely. Single artificial bee colony algorithm is easy to fall into localoptimal value. In order to overcome the defect, quantum artificial bee colony algorithmwas proposed by some scholars. The optimization results were superior to the simpleartificial bee colony algorithm. The quantum state in quantum artificial bee colonyalgorithm was described on plane unit circle in real area of Hilbert space. It has only avariable. The quantum properties did not give full play. It has difference to the idealeffect.This paper sets the continuous optimization problems for example. The quantumartificial bee colony algorithm based on Bloch coordinates of quantum bit was proposed.This algorithm uses Bloch coordinates of quantum bit encoding food sources in theartificial bee colony algorithm. The food source after encoding is corresponding to threefeasible solutions in optimization solution space. It makes the quantity of the globaloptimal solution get expansion.The new search algorithm improves the probability ofachieving the global optimal solution in the expanded space significantly. Food sourcesare updated by quantum rotation gate. This paper puts forward to a new method fordetermining the relationship between the two rotation phases in the quantum rotationgate. When the artificial bee colony algorithm searches as the equal area on the Blochsphere, it is proved that the size of the two rotation phases in the quantum rotation gateis approximate the inverse proportion. This avoids blind arbitrary rotation and makesthe search regular when approaching the optimal solution. The experiments of twotypical optimization issues show that the algorithm is superior to the common quantumartificial bee colony and the simple artificial bee colony in both search capability andoptimization efficiency. Finally combining quantum theory with neural network, ahybrid quantum neural networks model is put forward based on the relevant theoreticalknowledge of quantum neural network. Its validity is verified by two examples and theoptimization effect is better than that of common BP neural network.