When it comes to hard to solve problems, the significance of situational knowledge construction and network coordination must not be underrated. Professional deliberation is directed toward understanding, acting and analysis. We need smart and flexible ways to direct systems information from practice to network reflection, and to guide results from network consultation to practice. This article presents a case study proposal, as follow-up to a recent dissertation about online simulation gaming for youth care network exchange (Van Haaster, 2014).
Key to reinforcement learning in multi-agent systems is the ability to exploit the fact that agents only directly influence only a small subset of the other agents. Such loose couplings are often modelled using a graphical model: a coordination graph. Finding an (approximately) optimal joint action for a given coordination graph is therefore a central subroutine in cooperative multi-agent reinforcement learning (MARL). Much research in MARL focuses on how to gradually update the parameters of the coordination graph, whilst leaving the solving of the coordination graph up to a known typically exact and generic subroutine. However, exact methods { e.g., Variable Elimination { do not scale well, and generic methods do not exploit the MARL setting of gradually updating a coordination graph and recomputing the joint action to select. In this paper, we examine what happens if we use a heuristic method, i.e., local search, to select joint actions in MARL, and whether we can use outcome of this local search from a previous time-step to speed up and improve local search. We show empirically that by using local search, we can scale up to many agents and complex coordination graphs, and that by reusing joint actions from the previous time-step to initialise local search, we can both improve the quality of the joint actions found and the speed with which these joint actions are found.
LINK
This project builds upon a collaboration which has been established since 15 years in the field of social work between teachers and lecturers of Zuyd University, HU University and Elte University. Another network joining this project was CARe Europe, an NGO aimed at improving community care throughout Europe. Before the start of the project already HU University, Tallinn Mental Health Centre and Kwintes were participating in this network. In the course of several international meetings (e.g. CARe Europe conference in Prague in 2005, ENSACT conferences in Dubrovnik in 2009, and Brussels in April 2011, ESN conference in Brussels in March 2011), and many local meetings, it became clear that professionals in the social sector have difficulties to change current practices. There is a great need to develop new methods, which professionals can use to create community care.
MUSE supports the CIVITAS Community to increase its impact on urban mobility policy making and advance it to a higher level of knowledge, exchange, and sustainability.As the current Coordination and Support Action for the CIVITAS Initiative, MUSE primarily engages in support activities to boost the impact of CIVITAS Community activities on sustainable urban mobility policy. Its main objectives are to:- Act as a destination for knowledge developed by the CIVITAS Community over the past twenty years.- Expand and strengthen relationships between cities and stakeholders at all levels.- Support the enrichment of the wider urban mobility community by providing learning opportunities.Through these goals, the CIVITAS Initiative strives to support the mobility and transport goals of the European Commission, and in turn those in the European Green Deal.Breda University of Applied Sciences is the task leader of Task 7.3: Exploitation of the Mobility Educational Network and Task 7.4: Mobility Powered by Youth Facilitation.
