An abstract model for automated adaptable mobile agent

Abstract

A mobile agent (MA) typically operates in open, dynamic, and unpredictable environments. Therefore, when designing MA systems, it is difficult to anticipate all possible situations an MA may encounter. In order to reach its missions with intelligence capability, the MA must be able to adapt itself to the underlying environment. The goals of this study are to model adaptable MA and devise dynamic adaptation mechanisms necessary for the MA to operate in such environments by means of service discovery. In order to enable intelligence adaptation capabilities of the MA to accommodate the user's purpose, deterministic behavioral specifications are vital issues that must be addressed. Some scalable capabilities are applied toward the specifications so as to accommodate the bandwidth of its working environments. Two levels of adaptation are proposed. First, a deterministic gray box intention specification is proposed as a model which encapsulates the MA’s behavior in normalized form, operating with high available bandwidth. In which case, a detailed validation assessment is formulated to support a precision of matchmaking results. Second, a synopsis of deterministic intention specification is nominated in a compact form of MA's intention that must operate with low available bandwidth. A rough validation assessment corresponding with this synopsis analysis is presented. Accuracy of the results is measured by precision and recall as to how well the proposed model performs. The overall results indicate that this proposed model will permit the MA to be potent enough to accomplish its tasks within users' satisfaction.