Quantifying trust in computing networks

Abstract

Method for calculating a trust value of an agent in a computing network. In one implementation, the method may include receiving information pertaining to a first agent's previous actions, quantifying a discrepancy between an expected behavior and an actual behavior of the first agent during the first agent's previous actions, and determining the trust value of the first agent based on the quantified discrepancy.

1 . A method for calculating a trust value of an agent in a computing network, comprising:
 receiving information pertaining to a first agent's previous actions;   quantifying a discrepancy between an expected behavior and an actual behavior of the first agent during the first agent's previous actions; and   determining the trust value of the first agent based on the quantified discrepancy.   
     
     
         2 . The method of  claim 1 , wherein the discrepancy corresponds to a conditional entropy of the first agent based a second agent's experience in dealing with the first agent. 
     
     
         3 . The method of  claim 1 , wherein the discrepancy is characterized as a gap value. 
     
     
         4 . The method of  claim 1 , wherein quantifying the discrepancy comprises normalizing the discrepancy in an interval between 0 and 1. 
     
     
         5 . The method of  claim 4 , wherein determining the trust value of the first agent comprises subtracting the normalized discrepancy from 1. 
     
     
         6 . The method of  claim 1 , wherein the expected behavior comprises the first agent's promise to pay a specified amount in a previous transaction. 
     
     
         7 . The method of  claim 6 , wherein the actual behavior comprises an actual amount the first agent paid in the previous transaction. 
     
     
         8 . The method of  claim 1 , wherein the information pertaining to the first agent's previous actions comprises information regarding previous transactions provided by a credit card company. 
     
     
         9 . The method of  claim 1 , wherein the expected behavior comprises the first agent's promise to sell an item at a specified amount in a previous transaction. 
     
     
         10 . The method of  claim 9 , wherein the actual behavior comprises an actual amount the first agent sold the item in the previous transaction. 
     
     
         11 . A method for establishing trust groups in a computing network, comprising:
 receiving one or more trust values for each agent in a computing network;   identifying one or more agents having trust values that differ within a first specified tolerance; and   grouping the one or more agents into a trust group.   
     
     
         12 . The method of  claim 11 , wherein the trust group comprises a limit on the number of agents. 
     
     
         13 . The method of  claim 12 , further comprising splitting the trust group into sub-groups if the limit is exceeded. 
     
     
         14 . The method of  claim 11 , wherein the sub-groups comprise one or more agents having trust values that differ within a second specified tolerance. 
     
     
         15 . The method of  claim 11 , wherein the second specified tolerance is smaller than the first specified tolerance. 
     
     
         16 . A method for granting a new agent entry into a trust group within a computing network, comprising:
 receiving a request for entry into the trust group from the new agent;   sending information pertaining to the new agent to each member of the trust group;   receiving a trust value from each member of the trust group; and   forming a consensus of the trust values received from member of the trust group.   
     
     
         17 . The method of  claim 16 , further comprising: rejecting the request of the new agent for entry into one or more trust cliques if the consensus is below a predetermined value. 
     
     
         18 . The method of  claim 16 , further comprising: accepting the request of the new agent for entry into one or more trust cliques if the consensus is above a predetermined value. 
     
     
         19 . The method of  claim 16 , further comprising:
 splitting the trust group into two or more subgroups if the consensus is above a predetermined value and a limit on the number of members in the trust group has been exceeded; and   accepting the request of the new agent for entry into a subgroup.   
     
     
         20 . The method of  claim 16 , wherein the trust group is split based on one or more economic utility functions.