摘要

We combined the numeric genetic algorithm (NGA) and tabu search algorithm (TS) into a single tool (NGA-TS). Through computing a simulated system based on the relationship between the retention times and washing conditions, the NGA-TS can enhance the efficiency of the search algorithm and avoid entrapment into local optima. Computer-aided drug design has been an active research field in recent years. Our molecular docking program with NGA-TS as a search technique can be run on a PC. Empirical free energy values are used in the capacity of the scoring function. The performance of the program was verified by computing the beta-trypsin/benzamidine system. In addition, a test set of 100 protein-ligands complexes was studied as well. The results were compared with those of Autodock program and ConsDock. The proposed NGA-TS search methodology appears to be more efficient in terms of computational complexity (i.e. less memory and time consuming) than other routine search techniques. Therefore, it could be effectively applied to virtual screening of large databases.