İki boyutlu HP-model'inde protein katlanma problemine dağınık arama yaklaşımı

Proteins are complex molecules performing vital cellular activities. Each protein is composed of an amino acid chain. These amino acid chains are found as folded in their native state in the nature. The activity of a protein is mostly dened by its native 3 dimensional (3D) conformation. Determination of the native 3D conformation of an amino acid chain is called the protein folding problem (PFP). On the other hand, determination of the native state of a protein is not easy. In particular, protein folding problem is proven to be NP-hard. Determination of native states of protein molecules is vital since most of the medical and genetic studies depend on knowing these native conformations. Therefore many studies and investments are made on this topic. Researchers in various elds such as biochemistry, biophysics, computer science, operations research, bioinformatics, genetics and medical sciences are interested in solving this problem. Even, there are sometimes collaborations between dierent research areas to merge their techniques, knowledge and experience on this problem. One of these applications is the application of metaheuristics to protein folding problem in minimum free energy con- formation with HP-model based on a lattice structure, as a cooperation of all these sciences. This study is on developing a metaheuristic algorithm based on scatter search and path relinking for the 2D lattice based PFP utilizing HP-model according to minimum energy conformation theory. There are applications of well-known metaheuristics as well as scatter search to the PFP in the literature. We have introduced new moves and operators for the elements of scatter search for the PFP on 2D lattice HP-model. We have also proposed several application options for these element of the scatter search application, as well as new measurements for the values of the solutions. We introduced and experimented on some combinations of these elements for an eective search algorithm and give the test results in this study. Although the results are not as good as the results in the literature, we have tested a wide set of heuristic elements based on scatter search and path relinking. The results of these experiments are promising for further improvement on this topic.