Design Of Multi-objective Optimization Algorithm For DNA Coding

DNA computing is a parallel operation calculation model based on a large number of DNA molecules.It is a subject with great potential.Under the regulation of related enzymes,DNA molecules can spontaneously combine to react,producing a solution to the problem.High-quality DNA sequence molecules can prevent unnecessary non-specific hybridization,undesirable secondary structure and chemical instability in the DNA calculation process,resulting in failure of the reaction.DNA molecular design is a typical multi-objective optimization problem,which needs to meet various conflicting DNA coding target constraints.At present,many researchers at home and abroad have proposed many different DNA sequence design algorithms,but the DNA sequence constraint values designed by these algorithms are not uniform,and such sequences are still highly likely to undergo non-specific hybridization.Therefore,it is particularly important to be able to design DNA coding sequences with uniform and small constraint values.In this thesis,the reference point strategy is applied to the DNA coding problem,and a multi-objective DNA coding optimization algorithm based on the reference point is proposed.First,the weighted reference distance to the reference point is calculated according to each individual objective function value.Secondly,according to the reference distance from small to large in order,and set a crowded range ? for each individual to maintain the diversity of the population.Then,select enough individuals to perform selection,crossover and mutation operations to generate new populations.The DNA sequences of a group of 7 DNA sequences with a single strand length of 20 and a group of 14 DNA sequences with a single strand length of 20 were generated respectively,as well as their running time graphs,and the sequences were compared with the results of known literatures.Experimental results show that this algorithm can provide more reliable DNA sequences than existing sequence design techniques.In order to get better DNA sequence,combined with similar target merge and new balanced target,this paper proposes a multi-target DNA coding design method based on MMD algorithm.First,the Manhattan distance from each individual to the origin replaces the original fitness value.Second,select the best individual in the population,and update the fitness value of the original population according to the individual.Then,the current optimal individual is selected,and the fitness value of the original population is updated again until enough individuals are selected to generate a new population by using crossover and mutation operators.By comparing the generated DNA sequence with the results in the literature,the results show that the DNA sequence provided by the algorithm is more efficient and reliable.Finally,the convergence of the two algorithms proposed in this paper and the generated sequence are compared,and the second method is more excellent.