Meta-reasoning
Like planning, meta-reasoning contributes to enlarge rationality of intelligent agent in the contexts of performance quality (efficiency) as well as reasoning. Meta-reasoning, as the name implies, might consist of various meta-rules conjoining other capabilities such as learning and planning. For example, rational learning/planning requires deliberation about what to learn/plan about, when to lean/plan, why learn/plan it, to achieve optimal arrangement with limited resources, by multiple trade-off among conflicting factors involved in the current situation. During the trade-off of the conflicting factors, use of utility seems an orthogonal approach.
Issues in meta-reasoning
Architectures: