Behavioral agent testing of distributed information systems

Abstract

The use of multi-agent technologies in modern distributed information systems, that due to the increasingly demanded properties of autonomy, mobility, intelligence, cooperation, determines their appearance in the tools of online and offline testing. This work presents an agent-based check model of behavioral testing of a component of a distributed information system in its resource environment. This agent check model has the features of the composition of the reactive online and deliberative offline components of testing, suggesting deterministic and evolutionary methods of decomposition behavioral testing. To ensure autonomy and mobility, the agent-based check model defines the resource model of the component deployment environment, models of its goals and test strategies, signatures of multi-agent operations, as part of observation, strategy execution and adaptation, as well as initial models, goals and test strategies of the component. For the intelligence of the agent-based check model, its main components are also defined as individuals in the population of evolutionary test generation, which is proposed as an implementation of the deliberative component testing strategy. For the cooperation of agent-based check models in the network and multi-level interaction of their reactive and deliberative components, there are cooperative definitions for resource models of placement, goals and strategies, operations and initializations, realizability and transportability into the network DIS model, inheritance into the hierarchical DIS model. The check model defines the formal conditions/requirements for the performance of the behavioral online and offline testing, that performed by the multi-agent system of testing, and can be taken as the basic one, when constructing methods and special distributed behavioral test systems.