Single particle analysis is a field in computer science and biology, which tries to make 3D reconstruction of viruses from provided micrographs. Because some viruses are homogenous in nature (all of them may have same structure) so by applying some methods like ART or Weighted Back Projection and ... on provided images of single virus we can find 3D reconstruction of the virus in the computer. However there are some problems in this field. One of these problems arises when we have damaged particles (viruses) in the micrographs, this is very important because damage particles will produce not accurate reconstructions.
Position Guided Tabu Search:
Graph coloring is a very
interesting field of computer science. It is also very important because of its
important applications in the real world.
Tabu Search algorithm is a type of simulated annealing algorithm (algorithms that try to find global optimum of a function in a large space), these algorithms try to escape from local optima. Tabucol is the Tabu search algorithm for graph coloring. Tabucol have a problem in graph coloring. When we have a lot of local optima near each other then we may escape the first one and fall into the second one, we may escape second one and fall into third one and we may escape from third one and fall into first one. So because local search algorithms do not have memory then they may fall into a well full of local optimas and just go from one of them to another and never get a better result because these algorithms don't have memory. The Idea of PGTS (position guided Tabu Search) is to bring some kind of memory to Tabu search algorithm and try to escape from most of local optima wells.
PGTS have very good potential for upgrade, because the distance function of this algorithm is very simple and by bringing more accurate distance function we can improve the results (because we reduce error). I am trying to improve it.