| Studying the protein structure prediction problem is one of the frontier tasks in the field of life sciences. What it intends to solve is how to fold proteins into certain 3-dimensional structures from amino acid sequence of their primary structures. During the process of research, many simplified models have been proposed up to now for the sake of feasibility of computation. The corresponding protein structure prediction problems, however, are still concluded to be NP-hard problems. Consequently, it becomes more and more significant and demanding both in theory and in practicality to develop highly efficient approximate algorithm for resolving these problems.The quasi-human method is learning from human beings and their rich social experiences for wisdom to solve problem. The quasi-phsical method makes the original problem an optimization problem in mathematics. There is often the possibility of going to a local minimum of object function when solving the optimization problem mathematically. As for how to jump out of the trap of local minimum so that the calculation can head for a region with better prospects, the quasi-physical method is helpless. However, the quasi-human method can give us good strategies for jumping out of a trap of local minimum with the help of human beings' behavior and experiences.A new 3-dimensional real Euclid model is proposed for the protein structure prediction problem. Then a quasi-physical method based on corresponding physical model is presented to solve the relevant mathematical problem. Furthermore, an off-trap strategy of quasi-human approach is put forward in order to avoid frequently jumping into local minimum during the quasi-physical computation. Thus an efficient quasi-physical and quasi-human algorithm is achieved by integrating these strategie to solve this problem. Computational results indicated that the newly proposed model and its corresponding algorithm were both feasible and effective for the protein structure prediciton problem.
|
PDF Full Text Request