Context-Aware Task Sequence Planning for Autonomous Intelligent Systems

. The problem of context-aware task sequence planning by an autonomous intelligent system (intelligent agent) for a case of independent or loosely coupled tasks is considered. The principle of matching the task to the context was analyzed, the structure of the task sequence planning module and the algorithm of its work were proposed. The paper also proposes an algorithm for calculating the dynamic priority of a task, an algorithm for determining whether a task meets context, and an algorithm for adapting a set of rules for matching tasks to a context based on reinforcement learning in a stationary random environment with context dependence (contextual multi-armed bandit problem). The outline of the reinforcement learning procedure, implemented in the prototype of the task sequence planning module, has been proposed.