Search

In this work we present an empirical study where we demonstrate the possibility of developing an arti- ficial agent that is capable to autonomously explore an experimental scenario. During the exploration, the agent is able to discover and learn interesting options allowing to interact with the environment without any assigned task, and then abstract and re-use the acquired knowledge to solve the assigned tasks. We test the system in the so-called Treasure Game domain described in the recent literature and we empirically demonstrate that the discovered options can be abstracted in an probabilistic symbolic planning model (using the PPDDL language), which allowed the agent to generate symbolic plans to achieve extrinsic goals.

This paper describes how explicit resource reasoning is added within an existing temporal planning framework that uses timelines as plan representation. The work is grounded on a recent formalization of timeline-based planning that here is extended to model and reason over resources. The formal account is then fully implemented as an extension of the PLATINUm planner and tested to demonstrate its new capabilities.

Timeline-based planning is a paradigm that models temporal planning domains as sets of independent, but interacting, components. The behavior of the components can be described by means of a number of state variables whose evolution and interactions over time are governed by a set of temporal constraints. This paradigm is different from the one underlying the common action-based formalisms, such as PDDL, where the focus is on what can be done by an executive agent. Although successfully used in many real-world applications, little work has been done on the expressiveness and complexity of the timeline-based formalism. The present paper provides a characterization of the complexity of non-flexible timeline-based planning, by proving that a general formulation of the problem is EXPSPACE-complete. Such a result extends a previous work where the same complexity bound was proved for a restricted fragment of timeline-based planning that was shown to be expressive enough to capture action-based temporal planning. In addition, we prove that requiring an upper bound to the solution horizon as part of the input decreases the complexity of the problem, that becomes NEXPTIME-complete.

Timeline-based Planning has been successfully applied in several contexts to solve Planning and Scheduling (P&S) problems. A key enabling feature of the timeline-based approach to planning is its capability of dealing with temporal flexibility. Temporal flexibility is an important feature in real world scenarios. Indeed, it can be exploited by an executive system for robust on-line execution of flexible plans in order to absorb possible delays during the execution. In this regard, it is useful to define quality metrics to evaluate the robustness of flexible timeline-based plans. In this paper, a set of quality metrics for flexible timelinebased plans are defined and discussed when applied to a specific timeline-based framework. In fact, we consider the framework EPSL, developed to support the design of P&S applications, that allows the definition of different planners endowed with specific heuristics. Then, an experimental analysis is presented exploiting the planners to solve planning problem instances related to a real-world manufacturing case study. And, finally, an evaluation of planners performance is presented and discussed comparing results also considering robustness of generated plans.

One of the major problems facing our cities is the disposal of the huge amount of waste produced every day. A possible solution is represented by recycling. In this article, we propose a system for automatic recognition and extraction of materials from the unsorted waste, which takes advantage of Computer Vision and Machine Learning techniques. The system can classify the material of incoming objects and grasp, and insert them into proper bins. For the material classification phase, the system analyzes the information captured by a Near-Infrared (NIR) camera and an RGB camera. Experimental tests performed on real-world datasets show encouraging accuracy values.

This paper presents the planning system Pdk (Planning with Domain Knowledge), based on the translation of planning problems into Linear Time Logic theories, in such a way that finding solution plans is reduced to model search. The model search mechanism is based on temporal tableaux. The planning language accepted by the system allows one to specify extra problem dependent information, that can be of help both in reducing the search space and finding plans of better quality.

The “planning as satisfiability” approach for classical planning establishes a correspondence between planning problems and logical theories, and, consequently, between plans and models. This work proposes a similar framework for contingency planning: considering contingent planning problems where the sources of indeterminism are incomplete knowledge about the initial state, non-inertial fluents and non-deterministic actions, it shows how to encode such problems into Linear Time Logic. Exploiting the semantics of the logic, and the notion of conditioned model introduced in this work, a formal characterization is given of the notion of contingent plan (a plan together with the set of conditions that ensure its executability).