Multi-Agent Based Inter-Objects' Dynamic Relation Discovery and Service Recommendation in Interior Space

There are usually some dynamic relations between objects in interior space, which are often the true portrayal for human behavior, therefore translating inter-objects' dynamic relations into service sequence is imperative to serve people. At present, service accuracy improvement is mainly depended on collecting the related contexts from the interactions between human individuals and real-world objects to calculate inter-objects' dynamic correlation degrees for human individuals. But, a possibly correlated individual group appeared in the same interior space outside the targeted human individual is not account yet. To solve this problem, a three-phase method is proposed in this paper: (1) Simply abstract all objects into corresponding agents in order to eliminate the objects' nonuniformity and provide better services for human individuals; (2) Construct inter-objects' dynamic relation sequence from the existing context information due to the fact that most of the individuals' indoor behaviors are similar; (3) Evaluate the dynamic correlation degrees between the object-categories by mathematical expectation, so that more universal and accurate indoor services can be found and provided to wider population.