Multiple sequence alignment (MSA) represents a basic task for many bioinformatics applications. MSA allows finding common conserved regions among various sequences of proteins or DNA. However, to find the optimal multiple sequence alignment, it is necessary to design an efficient exploration approach that could explore a huge number of possible multiple sequence alignments. As well as, it is required to use a powerful evaluation method to assess the biological relevance of these multiple sequence alignment. To address these main problems, this article presents a multiobjective artificial fish swarm algorithm (MOAFS) to solve multiple sequence alignment. MOAFS uses the behaviors of artificial fish swarm algorithm such as the cooperation, decentralization and parallelism to ensure a good trade-off between the exploration and the exploitation of the search space of MSA problem. To preserve the quality and consistency of alignment, two fitness functions have been simultaneously used by the MOAFS algorithm: (i) Weighted Sum of Pairs to determine similar regions horizontally and (ii) Similarity function to determine vertically similar regions between the sequences of an alignment. Following the exploration of space search, the Pareto-optimal set is obtained by MOAFS which performs the optimal multiple sequence alignments for both fitness functions. The performance of MOAFS algorithm has been proved by comparing our algorithm with different progressive alignment methods, and other alignment methods based on evolutionary algorithms with single-objective and many-objective. The experiment results conducted on BAliBASE 2.0 and BAliBASE 3.0 benchmark confirm that the MOAFS algorithm provides a greater accuracy statistical significance in terms of SP or CS scores.

多序列比对（MSA）代表了许多生物信息学应用程序的一项基本任务。MSA允许在各种蛋白质或DNA序列之间找到共同的保守区。然而，为了找到最佳的多序列比对，有必要设计一种有效的探索方法，该方法可以探索大量可能的多序列比对。而且，需要使用功能强大的评估方法来评估这些多序列比对的生物学相关性。为了解决这些主要问题，本文提出了一种多目标人工鱼群算法（MOAFS）以解决多序列比对问题。MOAFS使用人工鱼群算法的行为，例如协作，分散化和并行性，以确保在探索和利用MSA问题的搜索空间之间取得良好的平衡。为了保持比对的质量和一致性，MOAFS算法同时使用了两个适应度函数：（i）对的加权和，用于水平确定相似区域；（ii）相似度函数，用于确定比对序列之间的垂直相似区域。在探索了空间搜索之后，通过MOAFS获得了帕累托最优集合，该集合对两个适应度函数都执行了最佳的多序列比对。通过将我们的算法与不同的渐进对齐方法以及基于单目标和多目标进化算法的其他对齐方法进行比较，证明了MOAFS算法的性能。