A Sequence Encryption Based On Multi-objective Differential Evolutionary Algorithm

As one of the core technologies in the field of cryptography, sequential cipher (alsoknown as stream cipher) technology is an important class of symmetric cryptosystem, thestrength of its core position is that each time it encrypts or decrypts the plaintext message(usually binary bits) only with a single character. The superiority of this technology is simple,fast and less error propagation. During the traditional methods of the sequential ciphergeneration, some tend to use a short key seed to produce a long cycle-sequence, or justconsider only a single constraint while ignore other randomness test calibrations. Suchsequential cipher is difficult to realize the true sense of the random sequence, but only meetspartially random sequence indicators. Response to this issue, this paper establishes amulti-objective optimization problem with the method of generating a sequence cipher, andproposes a sequence encryption based on multi-objective differential evolutionary algorithm(DEMOSEP) which can try to meet a number of the randomness indicators. This paper mainlyresearches the following three aspects:(1) Insight into the basic concepts of sequential cipher technology and several kinds ofclassical detection technology of sequential cipher algorithms, including RC4, A5, and SEAL.And shows the relevant principles of the graph, at the same time the advantages anddisadvantages of the three algorithms are pointed out. Then, discourses two populargeneration methods of the sequential cipher, one is based on the chaos, another is based on theintelligent computing.(2) The disadvantage of the traditional method of generating sequence cipher is onlyconsidering a single randomness indicator. Response to this issue, this paper establishes amulti-objective optimization problem with the method of generating a sequence cipher, and asequence encryption generating method based on multi-objective differential evolutionaryalgorithm (DEMOSEP) is presented. In this paper, three criteria used to evaluate therandomness of sequence are turned into three objective functions of multi-objectiveoptimization, and the algorithm uses the efficient and global search capability evolutionoperator of the differential evolutionary, moreover it merges the fast non-dominated sortingand crowded distance selection mechanism, thus evolves the stream ciphers to be satisfiedwith these three indicators. The experimental results show that the sequence encryption generating method has high randomness and security, and also has some practicality.(3) This paper researches the parameter settings of the DEMOSEP, the parameters in thestudy of evolution of cross probability and mutation probability in the calculation of the twokey control parameters were tested, and the appropriate parameter is obtained by thesimulation experiment to improve the DEMOSEP has better performance.