| Title | Towards Effective Genetic Trust Evaluation in Open Network |
| Publication Type | Conference Paper |
| Year of Publication | 2018 |
| Authors | Ma, S. |
| Conference Name | 2018 IEEE 20th International Conference on High Performance Computing and Communications; IEEE 16th International Conference on Smart City; IEEE 4th International Conference on Data Science and Systems (HPCC/SmartCity/DSS) |
| Keywords | behavior feedback, Biological cells, biological evolution process, crossover and mutation strategy, family attributes, genetic algorithm, genetic algorithm searching, genetic algorithms, genetic searching termination condition, genetic trust, genetic trust evaluation method, genetic trust search algorithm, Genetics, Human Behavior, human social trust propagation, human trust, intuitive trust, IP address, IP networks, malicious entities, misbehaving entities, Network security, open network environment, optimal trust chromosome, population evolutionary searching, pubcrawl, search problems, self-optimization, self-organization principle, self-organizing system, Sociology, Statistics, trust evaluation, trust evaluation process, Trust management, trust value, Trusted Computing |
| Abstract | In open network environments, since there is no centralized authority to monitor misbehaving entities, malicious entities can easily cause the degradation of the service quality. Trust has become an important factor to ensure network security, which can help entities to distinguish good partners from bad ones. In this paper, trust in open network environment is regarded as a self-organizing system, using self-organization principle of human social trust propagation, a genetic trust evaluation method with self-optimization and family attributes is proposed. In this method, factors of trust evaluation include time, IP, behavior feedback and intuitive trust. Data structure of access record table and trust record table are designed to store the relationship between ancestor nodes and descendant nodes. A genetic trust search algorithm is designed by simulating the biological evolution process. Based on trust information of the current node's ancestors, heuristics generate randomly chromosome populations, whose structure includes time, IP address, behavior feedback and intuitive trust. Then crossover and mutation strategy is used to make the population evolutionary searching. According to the genetic searching termination condition, the optimal trust chromosome in the population is selected, and trust value of the chromosome is computed, which is the node's genetic trust evaluation result. The simulation result shows that the genetic trust evaluation method is effective, and trust evaluation process of the current node can be regarded as the process of searching for optimal trust results from the ancestor nodes' information. With increasing of ancestor nodes' genetic trust information, the trust evaluation result from genetic algorithm searching is more accurate, which can effectively solve the joint fraud problem. |
| DOI | 10.1109/HPCC/SmartCity/DSS.2018.00105 |
| Citation Key | ma_towards_2018 |
Towards Effective Genetic Trust Evaluation in Open Network