An Investment-Rewards Model Based Paralleled Information Search Strategy

The purpose of this paper provides a investment and rewards based specific information search policy for large scale information network. An investment model and a reward model are introduced. The object of the models is to make the management agent to be able to instruct the search agents, even to form the information search activity more objective and controllable. The agents are organized into two groups, search agent group and management agent group. All agents can communicate to each other. These agents constitute a distributional search system network. A search agent determines their search paths based on the hyperlinks in the searched node, and the search reward of hyperlinks directed to. The most valuable node will be search after the reward evaluating to the objective target. An algorithm is given in the paper. From the results of the simulation experiment, it is obviously that the strategy is better than the other algorithms