在膜计算领域，一个备受关注的问题是证明各种膜系统模型能否在多项式时间内解决计算困难问题． 然而一个重要的事实是，许多能够求解 ＮＰ完全问题的膜系统模型甚至可以在多项式时间内解决 ＰＳＰＡＣＥ完全 问题，有的模型可以刻画 Ｐ＃Ｐ （该复杂类被推测为严格包含在 ＰＳＰＡＣＥ中）．文章主要介绍几类可以有效求解计 算困难问题的膜系统，包括活性膜膜系统、膜上带蛋白膜系统、组织膜系统、带膜分裂的同向／反向规则膜系统 以及脉冲神经膜系统；概述了这些膜系统模型的计算复杂性，指出了一些可以提高膜系统计算能力的特性，最 后给出膜系统中存在的一些公开问题．

In the area of membrane computing, a great attention is traditionally paid to how to demonstrate a theoretical possibility to solve hard computational problems in polynomial time by means of various P system models. However, an important fact is that some P system models with this capability are actually stronger: they are able to solve PSPACE complete problems in polynomial time, while others have been recently shown to characterize the class P#P , which is conjectured to be strictly included in PSPACE. In this survey, we focus on some P system models which have the potential to solve hard computational problems, including P systems with active membranes, tissue P systems, P systems with proteins on membranes, P systems with symport / antiport and membrane division, as well as spiking neural P systems. We provide a survey of computational complexity results of these P system models, point out some features and provid them with their computational strength. Finally, some open problems related to the above mentioned P systems are presented.