Research On Optimization Algorithms Based On Quantum Computing Principles

Quantum computing is seen as one of the most efficient means to enhance computingcapacity for the current physical system, since its incomparable superiority. As an importantmeans of implementing quantum computing, the optimization algorithm based on quantumcomputing principles has become an important research interest of intelligent computing. Itinvolves not only the pure quantum searching algorithm based on the principles of quantumand quantum gate circuit, but also the quantum inspiring intelligent algorithm combined thetraditional intelligent optimization algorithms with the quantum computing mechanism,becomes a hot and key topic in the current optimization research fields. In this paper, sometypical quantum optimization algorithms are intensively studied, the structural characteristicsof these algorithms and their problems in application are summarized, and some improvementschemes are proposed. The innovative results are as follows:(1) A Grover quantum search algorithm based on fixed targets weight is proposed to deal withthe inability of the Grover search algorithm in identifying the difference among targets. Theinitial superposition states are constructed according to the difference of importance fortargets, and the unitarity of operator is proved, then the properties of the improved scheme arededuced. On this basis, the GRK quantum partial searching algorithm based on fixed targetsweight is also deduced. The simulation experiments show that both algorithms can find thetargets successfully according to the assigned weight value. Finally, a controllable quantumsequential multi-signature scheme based on the improved GRK algorithm is proposed.(2) Quantum cloning immune algorithm based on cloud model is proposed. Quantum rotationgate, the key operator in quantum evolutionary algorithm, is replaced by cloud modelcollaborative operators; the selective population cloning is realized according to chromosomeindividual fitness, significantly improve the convergence speed and global search ability ofevolutionary algorithm. On this basis, another improved algorithm based on qubits phasecoding is proposed, and can be applied to continuous space optimization effectively. Thesimulation shows the algorithm can make accurate estimation value for the nonlinear system.(3) Quantum particle swarm optimization algorithm based on cloud model is put forward. Itgreatly improves the performance of the algorithm only through self-adaptive control thecontraction-expansion factors by cloud model operator while keeping the simplicity andspeediness of quantum particle swarm algorithm. A binary quantum particle swarm algorithmfor discrete space optimization is also proposed according to quantum well model. Based onthe above two algorithms, three kinds of compressed sensing signal reconstruction schemes are put forward, all of which achieve satisfactory effects of signal reconstruction.(4) Three kinds of quantum ant colony optimization algorithms are proposed to alleviate weakperformance of ant colony algorithm in continuous space optimization and TSP. By adaptivecontrol for the angle of quantum rotation gate and pheromone of ant colony, the performanceis improved effectively in continuous space, especially in high dimensional space; theprobability choice model and pheromone update model are redefined integrated with quantuminformation intensity factors, and dynamic adjustment for parameters is also strengthened, sothe performance of the algorithm in solving TSP is greatly improved; Cloud model isintegrated in sampling process of the Gaussian kernel function, greatly improve theconvergence speed and global search ability of extended ant colony algorithm. On this basis,the improved algorithm is integrated with BP algorithm in neural network, further improvesthe accuracy of signal recognition.