To search for a target n-product Boolean vector of fixed weight d, we propose an important method involving the notion of a fixed-weight "vector label" accompanied with a vector label restoration algorithm. Based on these, we present a new quantum algorithm designed to search for a fixed-weight target whose computation complexity, specifically ( ) 1 d n O C + , is better than that for a classical algorithm. Finally, we use the procedure to search for the NTRU private key as an example to verify the efficiency of the new algorithm in searching for fixed-weight target solutions.
label, quantum search, computation complexity, NTRUCitation: Wang X, Bao W S, Fu X Q. A quantum algorithm for searching a target solution of fixed weight. Quantum computation evolved from proposals put forward by Paul [1] and Feynman [2] to harness the computational power of quantum environments. In 1992, Deutsch et al. [3] designed the first quantum algorithm whose computation capability surpassed that of an electronic computer. Shortly thereafter, Shor's [4] quantum factorization algorithm and Grover's [5] quantum search algorithm added further to the security difficulties associated with modern cryptology.Since then, to overcome this, scholars [6][7][8][9] in China and abroad have paid great attention to research into quantum computer technology, quantum computation algorithms and quantum cryptology. Grover's algorithm, which offers an exhaustive search of long lists held in a quantum computer, reduces the computational complexity of current exhaustive search attacks from O(2 n ) to O(2 n/2 ), thus raising the security concerns of current crypto-systems to a new level. Along with an in-depth study, Long et al. [10][11][12][13] presented an improved algorithm based on phase shifts with zero theoretical failure rate, while Zhong et al. [14] proposed a meet-in-the-middle quantum search algorithm. As a result, Grover's algorithm *Corresponding author (email: bws2004@sina.com) has been widely applied.Indeed, Grover's algorithm is a universally applicable algorithm for which the input is an equally-weighted superimposed state of all classical n-product states, each of which is potentially a target solution. However, in practical decryption of a cryptosystem, such problems can be represented as searches for a target solution with special conditions (as a target-product vector with fixed weight). If Grover's algorithm is used to find a solution, then the computation complexity is no better than for classical search algorithms, sometimes even much worse. Reasoning in this way, the exploitation of a quantum computer's powerful paralleling capability seems not to be fully achieved. Obviously, improving the efficiency of classical crypto-analysis with the use of quantum computation principles is a key and vital problem needing to be resolved. In this article, we focus on solving this special crypto attack problem associated with target vectors of fixed weight, and we propose an important method introducing the notion of a fixed weight vector labe...