In this paper, we consider the problem of solving a large MDP in a distributed way among several processors. To do that, we propose an approach which decomposes the large MDP into smaller ones each of which is solved on a unique processor. The obtained joint local policies derived from the small MDPs (subMDPs) behave in the same way of the policy of the initial MDP.
This work pursues an empirical study on the mutual interactions among a set of off-line and on-line constraint-based scheduling approaches. We devise a set of closed loop execution management algorithms, and compare their behavior within an experimental framework which allows to directly assess the consequences of each chosen strategy combination against the injection of a number of disturbing events, through simulated schedule executions.
Peter Votruba, Andreas Seyfang, Michael Paesold, Silvia Miksch
847 - 848
An important application of both data abstraction and plan execution is the execution of clinical guidelines and protocols (CGP), both to validate them against a large set of test cases and to provide decision support at the point of care. CGPs can be represented and executed as a hierarchy of skeletal plans. To bridge the gap between low-level data and high-level concepts in the CGP, intelligent temporal data abstraction must be integrated with plan execution.
In this paper we describe a solution to this challenge which was implemented as part of the European project Protocure II to improve the quality of CGPs. They are translated to the high-level plan representation language Asbru which again is compiled into a network of abstraction modules by the system. Then this network performs the content of the plans triggered by the arriving patient data.
By this, we seamlessly integrate the synchronisation of guideline execution with observed patient state, complex temporal abstractions and execution of complex plans without requiring the user to handle the low-level details. Instead, user-friendly tools are used to create and maintain the guideline.
In this paper, we present a multiagent model in which agents have a perception upon their shared environment, a measure is associated to the agents' perception field. We apply the model on the Vehicle Routing Problem with Time Windows (VRPTW). The overall process adopts the general schema of parallel insertion methods and it uses the contracting of perception's field of the vehicle agents as a new distance between them. This new measure expresses the feasibility universe of the vehicles and is used as a criterion of choice between candidates vehicles for the insertion of a customer in their plan. Our approach provides a new method to tackle the Time constrained VRP in which the solving process is focused on the future and constitutes an alternative for handling the dynamic version of the problem.
Designing and implementing an intelligent system that tackles the problem of placing two-dimensional shapes on a surface, such that no shapes overlap and the uncovered surface area is minimized, is highly important in industrial applications. However, it is also interesting from the scientific perspective, in terms of artificial intelligence, since autonomous systems developed up to now have found it difficult to compete with humans in this task. This paper presents a new algorithm which addresses the on-line packing of two-dimensional irregular shapes, and achieves high quality solutions in short computational times. The key point of this algorithm is the utilization of techniques drawn from computer vision and artificial intelligence.
In this paper we present a mobile robot localization system that integrates Monte-Carlo localization with an active action-selection approach based on an aliasing map.The main novelty of the approach is in the off-line evaluation of the perceptual aliasing of the environment and in the use of this knowledge to perform localization processes faster and better. Preliminary results show improved performances compared with the classic Monte-Carlo localization approach.
The emerging field of ubiquitous robotics presents new challenges for human-robot interface. In this note, we introduce the concept of a common interface point using an expression-based semantics as a way to address some of these challenges. We illustrate this concept in the framework of the PEIS-Ecology approach to ubiquitous robotics.
In this paper, we present a leaf classification method based on skeletons produced by a navigation-inspired technique. The classification system comprises three separate stages. First, a
skeletonisation algorithm is used to gather low level structural and morphological information about the shape. Subsequently, the data is converted into a series of attributed graphs. Graphs of the same type are then compared using an approximate graph matcher, which identifies a degree of similarity between them. Each degree of similarity corresponds to a dimension in a conceptual space, as defined by Gärdenfors. We test the performance of our technique on a set of leaves belonging to three different species.