Planning with trial and errors

Agent-based systems have gained immense popularity in the last decade or so. Reasoning methods such as planning and decision-making under uncertainty applied to agents and multi-agent systems is an important area of research in agent-based systems. Most works in planning in uncertain environments incorporate sensing actions. We wish to address planning problems when there are no such sensing actions. For this, we introduce the notion of dasiatry actionspsila and develop a logic PDL_try that is an extension to propositional dynamic logic (PDL). The logic also allows specification of the implicit knowledge of an agent. We illustrate planning in our logic with a new adaptation of the popular blocks world domain. Example programs (plans) are discussed in detail for single agent and two agent situations.