A research plan for deri innsbruck movin

DERI INNSBRUCK Leopold-Franzens Universität Innsbruck DERI – Digital Enterprise Research Institute A Research Plan for DERI Innsbruck: Moving from software to serviceware and from syntax to semantics Jos de Bruijn, Alice Carpentier, Ying Ding, Dieter Fensel, Martin Hepp, Stijn Heymans, Holger Lausen, Birgit Leiter, Christian Mayer, Melanie Plattner, Thomas Strang, Michal Zaremba DERI Galway National University of Ireland Galway Ireland www.deri.ie DERI Innsbruck University of Innsbruck Technikerstrasse 21a Innsbruck Austria www.deri.at DERI Korea 267 Deokil Bldg (Saltlux), Daechi-dong, Gangnam-gu, Seoul 135-848 Korea www.deri-korea.org DERI Stanford Stanford University Serra Mall Stanford USA www.deri.us September 20, 2006 Abstract A large research body needs a structure to facilitate the potential strength implicitly present in its size This report is about releasing the full potential that DERI Innsbruck has in this respect We derive objectives from the overall vision of DERI and align them with researchers and research projects through the means of research cluster INTRODUCTION SURVEY 2.1 OBJECTIVES 2.2 CLUSTERS 14 2.3 PROJECTS 15 2.4 RESEARCH BODY .17 2.4.1 Student researchers 17 2.4.2 Junior researchers 17 2.4.3 Senior researchers 24 REASONABLE SEMANTIC WEB SERVICES CLUSTER (RSWS) .25 3.1 GENERAL DESCRIPTION 25 3.2 OBJECTIVES .27 3.2.1 Discovery 27 3.2.2 Choreography 29 3.2.3 Formal Languages .32 3.2.4 Reasoning 38 3.3 PROJECTS 42 3.3.1 Infrawebs 42 3.3.2 Knowledge Web 44 3.3.3 RW² 45 3.3.4 Salero .47 3.3.5 Sekt 48 3.3.6 SemNetMan .49 3.3.7 SenSE 50 3.4 STAFF 51 3.4.1 Student Researchers 52 3.4.2 Junior Researchers 52 3.4.2.1 Darko Anicic 53 3.4.2.2 Jos de Bruijn 56 3.4.2.3 Dimitrij Denissenko 63 3.4.2.4 Cristina Feier 64 3.4.2.5 Uwe Keller 68 3.4.2.6 Holger Lausen 73 3.4.2.7 Ruzica Piskac 78 3.4.2.8 Richard Pöttler 80 3.4.2.9 James Scicluna 81 3.4.2.10 Alexander Wahler 85 3.4.3 Senior Researchers 87 3.4.3.1 Stijn Heymans 87 SEMANTICS IN BUSINESS INFORMATION SYSTEMS CLUSTER (SEBIS) .92 4.1 GENERAL DESCRIPTION 92 4.2 OBJECTIVES .94 4.2.1 Ontologies 94 4.3 PROJECTS 99 4.3.1 DIP 99 4.3.2 EASAIER 101 4.3.3 EastWeb 102 4.3.4 EnIRaf .103 4.3.5 etPlanner 104 4.3.6 MUSING 105 4.3.7 myOntology 106 4.3.8 OnTourism 108 4.3.9 SUPERMartin .110 4.4 STAFF .112 4.4.1 Student Researchers 112 4.4.2 Junior Researchers 112 4.4.2.1 Tobias Bürger 113 4.4.2.2 Jan Henke 116 4.4.2.4 Dumitru Roman 117 4.4.2.5 Francois Scharffe 122 4.4.2.6 Katharina Siorpaes 125 4.4.2.7 Michael Stollberg 126 4.4.3 Senior Researchers 131 4.4.3.1 Ying Ding 131 4.4.3.2 Martin Hepp 137 SEMANTIC EXECUTION ENVIRONMENT CLUSTER (SEE) 144 5.1 GENERAL DESCRIPTION 144 5.2 OBJECTIVES 147 5.2.1 Applications 147 5.2.2 Developer tools 149 5.2.3 Mediation .153 5.2.4 Execution management 156 5.3 PROJECTS 159 5.3.1 Adaptive Service Grid 159 5.3.3 SEEMP 160 5.3.4 SemanticGov 161 5.3.5 SemBiz 162 5.3.6 SUPERMichal .163 5.3.7 TSC 165 5.4 STAFF .167 5.4.1 Student Researchers 167 5.4.2 Junior Researchers 167 5.4.2.1 Emilia Cimpian 168 5.4.2.2 Graham Hench 171 5.4.2.3 Zhou Jingtao 173 5.4.2.4 Mick Kerrigan 180 5.4.2.5 Adrian Mocan 185 5.4.2.6 Omair Shafiq 190 5.4.2.7 Adina Sirbu 196 5.4.2.8 Zhixian Yan 198 5.4.3 Senior Researchers 199 5.4.3.1 Michal Zaremba 199 UBIQUITOUS SERVICES CLUSTER (UBISERV) 206 6.1 GENERAL DESCRIPTION 206 6.2 OBJECTIVES 207 6.2.1 Adaptation .207 6.2.2 Grounding .209 6.2.3 Storage & Communication .210 6.3 PROJECTS 215 6.3.1 GRISINO .215 6.3.2 SWING 217 6.3.3 TripCom 218 6.4 STAFF .220 6.4.1 Student Researchers 220 6.4.2 Junior Researchers 220 6.4.2.1 Jacek Kopecky 220 6.4.2.2 Reto Krummenacher 223 6.4.2.3 Ioan Toma 228 6.4.3 Senior Researcher 236 6.4.3.1 Michael Jäger 236 6.4.3.2 Axel Polleres 236 6.4.3.3 Thomas Strang 238 BEYOND RESEARCH: TEACHING, BUSINESS DEVELOPMENT, AND GENERAL MANGEMENT 246 7.1 TEACHING 246 7.2 BUSINESS DEVELOPMENT 251 7.2.1 General Description 251 7.2.2 Projects 252 7.2.2.1 DERIBusinessDevelopment 253 7.3 CENTRAL MANAGEMENT UNIT 255 7.3.1 DERIExchange .255 7.3.2 DERISustainability 255 7.3.3 Knowledge Web Network 256 REFERENCES 257 APPENDIX 262 Introduction In general, a research institutes based on external funding has three major challenges to meet: • • • It needs to provide excellent research results to justify its existence It needs to provide excellent education for its researchers to mature its outcomes It needs to provide excellent performance in research projects to ensure its funding Unfortunately, these three dimensions may define conflicting requirements In consequence it is essential to align them properly We have chosen a top-down approach where an overall vision and mission is used to align these dimensions properly In [1], the vision of serviceware as the next natural step beyond hardware and software is introduced: “After four decades of rapid advances in computing, we are embarking on the greatest leap forward in computing that includes revolutionary changes at all levels of computing from the hardware through the middleware and infrastructure to applications and more importantly in intelligence This paper outlines a comprehensive framework that ingtegrates two complimentary and revolutionary technical advances, ServiceOriented Architectures (SOA) and Semantic Web, into a single computing architecture, that we call Semantically Enabled Service-Oriented Architecture (SESA) While SOA is widely acknowledged for its potential to revolutionize the world of computing, that success depends on resolving two fundamental challenges that SOA does not address, integration, and search or mediation In a services-oriented world, billions of services must be discovered and selected based on requirements, then orchestrated and adapted or integrated SOA depends on but does not address either search or integration The contribution of this paper is to provide the semantics-based solution to search and integration that will enable the SOA revolution The paper provides a vision of the future enabled by our framework that places computing and programming at the services layer and places the real goal of computing, problem solving, in the hands of end users.” Based on this SESA vision, a top down approach for organizing the research body in Innsbruck is developed This implies the following: • • • Objectives are derived from the mission to realize the SESA vision Projects must contribute to one or several of the components of SESA, probably on a 80% rule, i.e., 20% can be about different or related topics This reflects the need for opportunisms since we are cooperating in this process with funding agencies and external partners that both have their own agendas Researcher and their research topics follow from sub aspects of some of the objectives Again, this should hold for at least 80% of these topics Finally, DERI Innsbruck has clusters as a means to decompose research, the large number of researchers, and project responsibility Each cluster is responsible for a number of objectives, a number of researchers, and a number of projects In the following section, namely Section 2, we provide a general summary on goal, project, and research staff distribution over clusters The subsequent sections, Section to Section 6, introduce the four research clusters of DERI Each of these sections is divided into the following subsections, general description, objectives, projects and staff Section adds further activities of DERI beyond its focus on research These activities are related to teaching, business development, and general management Survey We will survey objectives, clusters, projects and research staff 2.1 Objectives A objective usually combines a research area, i.e., a major research challenges in SW(S) and SESA together with an implementation effort related to it.1 A objective typically has a corresponding architectural component, and vice versa A tight coupling between objectives and architectural components is desirable The WSMX platform [2] provides a SESA environment which facilitates prototype development The major outcome of each of the objectives consists of peer-reviewed conference and journal publications The prototypes associated with the papers are the major outcome of the architectural components which are associated with the research components We distinguish different types of elements of an overall SESA where each element type is composed by some sub functionalities: • • • • The problem-solving layer which consists of (1) Ontologies, (2) Applications (e.g., e-tourism, e-government) and (3) Developer tools (GUI tools such as ontology/web service description engineering tools; generic developer tools such as language APIs, parsers/serializers, converters, etc.) The broker layer which consists of (4) Discovery, (5) Adaptation (including selection and negotiation), (6) Composition (web service composition techniques such as planning), (7) Choreography, (8) Mediation ((a) Ontology mediation: techniques for combining Ontologies and for overcoming differences between Ontologies; (b) Process mediation: overcoming differences in message ordering, etc.), (9) Grounding, (10) Fault Handling (Transactionality, Compensation, etc.), and (11) Monitoring The base layer that is providing the exchange formalism used by the architecture, i.e., (12) Formal languages (static ontology and behavioral, i.e., capability/choreography/orchestration languages, connection between higherlevel descriptions, e.g., WSML), (13) Reasoning (techniques for reasoning over formal descriptions; LP, DL, FOL, behavioral languages, etc.) and (14) Storage and Communication Finally, vertical services such as (15) Execution management and (16) Security (authentication/authorization, encryption, trust/certification) Existing working group such as WSML will become a working group of a certain objective The following image presents the current status of WSMX architecture Figure 2.1.1 SESA Architecture Hereby, the overall roadmap is as following: • • Currently, DERI Innsbruck focuses on the following essential components to boot-strap the overall approach: (1) Ontologies, (2) Applications, (3) Developer tools, (4) Discovery, (5) Adaptation, (6) Composition, (7) Choreography, (8) Mediation, (9) Grounding, (12) Formal languages, (13) Reasoning, (14) Storage and Communication, (15) Execution management There are no concrete plans yet for (10) Fault Handling, (11) Monitoring, and (16) Security Some of this work may be provided by external DERI cooperation partners The following table summarizes these objectives and their leaders Objectives No Objective Cluster SEBIS Ontologies In this research topic, we want to advance the state of the art in the creation and the use of ontologies for the automation of business processes Ontologies in our understanding are community contracts about a representation of a domain of discourse Representation in here includes (1) formal parts that can be used for machine reasoning, and (2) informal parts like natural language descriptions and multimedia elements that help humans establish, maintain, and renew consensus about the meaning of concepts Our research output will Leader Martin Hepp [22] A Mocan, E Cimpian, M Zaremba, C Bussler: Mediation in Web Service Modeling Execution Environment (WSMX), Information Integration on the Web 2(iiWeb2004), August 2004, Toronto, Canada [23] A Mocan, M Zaremba; Mediation in Web Service Modeling Execution Environment (WSMX), Information Integration on the Web Workshop (iiWeb2004), Conference on Very Large Data Bases VLDB2004, August 2004 [24] C Bussler, E Cimpian, D Fensel, J M Gomez, A Haller, T Haselwanter, M Kerrigan, A Mocan, M Moran, E Oren, B Sapkota, I Toma, J Viskova, T Vitvar, M Zaremba, M Zaremba: Web Service Execution Environment (WSMX) W3C Member Submission April 2005 [25] M Zaremba, M Moran, T Haselwanter: Semantic Web Services Architecture and Information Model, OASIS Semantic Execution Environment (SEE) Technical Committee, Working Draft, March 2006 (in progress) [26] J Domingue, B Norton, O Shafiq, M Zaremba: Semantic Execution Environment (SEE) Execution Semantics, OASIS Semantic Execution Environment (SEE) Technical Committee Working Draft, March 2006 (in progress) [27] A Haller, M Zaremba, Mission Statement - WSMX, D7.3v1.0 WSMX Working Draft January 2005 [28] E Cimpian, M Moran, E Oren, T Vitvar, M Zaremba: D13.0 Overview and Scope of WSMX [29] E Cimpian, A Mocan, M Moran, E Oren, M zaremba: D13.1 Web Service Execution Environment - Conceptual Model (WSMX_O) [30] A Mocan, E Cimpian: D13.3 WSMX Data Mediation [31] E Cimpian, A Mocan: D13.7 Process Mediation in WSMX [32] Michal Zaremba, Matthew Moran, Thomas Haselwanter, WSMX Architecture, D13.4v0.2 WSMX Working Draft [33] Matthew Moran, Adrian Mocan, Michal Zaremba, Integration of WSMX with other SWS Systems, D13.8v0.1 - WSMX Final Draft 14-06-2005 [34] D Aiken, M Zaremba: WSMX Documentation, D22.0v0.2 WSMO Working Draft, February 2005 [35] Mick Kerrigan: D9.3v0.1 WSMX Monitor, WSMX Working Draft, January 2005 [36] Mick Kerrigan: D9.4v0.1 WSMX Invoker, WSMX Working Draft, June 2005 259 [37] A Haller: D13.6v0.1 WSMX Use Cases, WSMX Working Draft, December 2004 [38] A Haller, J Scicluna, T Haselwanter: D13.9v0.1 WSMX Choreography, WSMX Working Draft, June 2005 [39] E Kilgarriff: D13.12v0.1 WSMX Discovery Component, WSMX Draft November 2005 [40] J Kopecký, D Roman: D24.1v0.1 Aligning WSMO and WSMX with existing Web Services specifications, WSMO Working Draft January 2005 [41] C Bussler, E Kilgarriff, R Krummenacher, F Martin-Recuerda, I Toma, B Sapkota: D21.v0.1 WSMX Triple-Space Computing, WSMO Working Draft June 2005 [42] D07.02: The Web Service Execution Environment: Moran, M., Zaremba, Michal, Haselwanter, T., Zaremba, Maciej, Oren, E., 2005 [43] D14.02: WSMX Interoperability with Related Semantic Web Service Systems: Moran, M., Zaremba, Michal, Mocan, A., 2005 [44] D07.01a: Version of the Web Service Design Environment: Moran M., Kilgarriff E.,HallerA.,June2004.Availableat:http://lion.deri.ie/PDFdeliverables/D07.01a_v0.0 1.html [45] D07.03a v0.1: Web Service Execution Environment, Moran M., Zaremba M., Oren E.,MocanA.,CimpianE.June2004.Availableat:http://lion.deri.ie/PDFdeliverables/D0 7.03a_v0.01.html [46] D07.09a: Integrated Web Service Design and Execution Environment, Moran M June 2004 Available at: http://lion.deri.ie/PDFdeliverables/D07.09a_v0.01.html [47] A Mocan, E Cimpian: D13.3v0.2 WSMX Data Mediation, WSMX, 2005 [48] E Cimpian, A Mocan, D13.7 v0.1 Process Mediation in WSMX , WSMX, 2005 [49] A Mocan, E Cimpian D13.3v0.1 WSMX Mediation, WSMX, 2004 [50] E Cimpian, A Mocan, M Moran, E Oren, M Zaremba [51] D13.1v0.1 WSMX Conceptual Model, WSMO, 2004 260 [52] Michal Zaremba, Matthew Moran, Emilia Cimpian, Adrian Mocan, Eyal Oren D13.5v0.1 WSMX Implementation, WSMO, 2004 [53] E, Cimpian, A Mocan: D5.1b Process Mediation - Concepts, Architecture and Implementation in WSMX, June 2005 [54] A Mocan, E, Cimpian: D5.1a Data Mediation - Concepts, Architecture and Implementation in WSMX, June 2005 [55] A Mocan, E, Cimpian: D5.01a Initial Report on Web Service Mediation Framework, June 204 [56] WP6 Interoperability & Architecture, D6.5 DIP Revised Architecture: Zaremba Michal, Moran, M., Haselwanter, T., Zaremba, Maciej, Oren, E, 2005 [57] WP6 Interoperability & Architecture, D6.2 DIP Architecture: Hauswirth M., Schmidt R., Altenhofen M., Drumm C., Bussler C., Moran M., Zaremba M., Vasiliu L., Quantz J., Henocque L., Haller A., Sapkota B., Kilgarriff E., Petkov S., Aiken D., Oren E., Ohlendorf M., Mocan A December2004 Availableat:http://dip.semanticweb.org/documents/D6.2-DIP-Architecture.pdf [58] E Cimpian, M Zaremba, B Sapkota, J Domingue, L Cabral: D5.5 Business Data and Process-Level Mediation Module Prototype v2, January 2006 [59] E Cimpian, J Lemcke, A Mocan, M Schumacher: D5.3a Business Process-level Mediation Module Specification, June 2005 [60] S Arroyo, E Cimpian, M Dimitrov, J Domingue, G Nagypal, M Spork, L Vasiliu: D3.2 Service Description Framework, June 2004 [61] E Cimpian, C Drumm, M Stollberg, I Constantinescu, L Cabral, J Domingue, F Hakimpour, A Kiryakov: D5.1Report on the State-of-the-Art and Requirements Analysis (WP - Service Mediation), June 2004 [62] O Shafiq, R Krummenacher, Y Ding, B Draxler: D4.1 Integration of WSMX and Triple Space Computing Architectures, TSC Project Deliverable (in-progress) [63] A Hevner, S March, J Park, and S Ram Design Science in Information Systems Research MIS Quaterly, 28(1):75–105, 2004 261 Appendix Not yet assigned components: Nr Title Mission statement Web site Leader Cluster Team Composition Develop methods to web service composition (WSC), starting from web service descriptions at various levels of abstraction, specifically, the functional level and process level components of WSMO Implement such methods as tools in the relevant contexts, in particular WSMX Find potential applications of WSC technology, model them using WSMO/WSML, and run case studies with the developed tools, ultimately resulting in technology export Jörg Hoffmann Senior Researchers: Jörg Hoffmann Junior Researchers: Students: - Contributing projects Since the working group has yet to be started in Sep'06 – the current Current status mainly comes down to the previous work done by Jörg Hoffmann Status The relevant work was done in the area of AI Planning, in particular in Planning under uncertainty The plan for the first year of the working group is: Sept'06 - Dec'06: Build on Jörg Hoffmann's previous work to obtain a prototype doing functional level composition 22 October 06 – August 07: Contact DERI collaborators from the industry to solicit potential applications of WSC technology Model the applications in WSMO/WSML, and run case studies with the developed technology January 07-August 08: Turn the prototype into a useable and reasonably functional WSMX tool January 07-August 08: Start to develop methods for process level composition, in close collaboration with IRST Trento A preliminary step towards is currently being undertaken, in collaboration with IRST Trento, in the context of the KnowledgeWeb WSC prototype deliverable due end Jun'06 In what follows, we briefly sketch Jörg Hoffmann's relevant previous work, and the work done for the KnowledgeWeb prototype We then fill in a few more details on steps - In AI Planning, "actions" must be composed into "plans" The formal framework are declarative transition systems: given a vector of state variables, a set of transition rules (the actions), a start state (value assignment), and a target condition, find a sequence of transitions (a plan) that leads from the start state to a state that satisfies the goal Actions are described in terms of their precondition and effects, both of which are simple, mostly formulas over the state variables At an abstract level, this is quite similar to WSMO functional level service descriptions; more below Jörg Hoffmann developed the "FF" planning system, that revolutionized AI Planning when it excelled in the international planning competitions 2000 and 2002 [Hoffmann&Nebel JAIR'01, Hoffmann ECAI'02, Hoffmann JAIR'03] More recently, FF variants "Conformant-FF" [Brafman&Hoffmann ICAPS'04, Hoffmann&Brafman AI'06], "Contingent-FF" [Hoffmann&Brafman ICAPS'05], and "ProbabilisticFF" [Domshlak&Hoffmann ICAPS'06] have been developed, dealing with "conformant", "contingent", and "probabilistic" planning, respectively In conformant planning, instead of a start state one has a description in the form of a formula of a set of possible start states, where the plan must work for every possible start state Contingent planning extends this with special observation actions that correspond to "if" statements to be carried out at plan execution time, so that a plan is now a tree no longer a sequence of actions In probabilistic planning, finally, the "start state" is a probability distribution in the form of a Bayesian network over states, and the plan must achieve the goal with a probability greater than a given threshold In WSC, there are various sources of uncertainty for example, about what kind of instance an input will be instantiated with, see below and it is reasonable to assume that (some of) the techniques used in Conformant-FF and Contingent-FF will help to deal with this uncertainty Other work of Jörg Hoffmann that is likely to become Relevant for WSC (see also Future Steps below) is, e.g., his work on problem structure [Hoffmann AIPS'02, Hoffmann JAIR'05, Hoffmann et al ICAPS'06a], abstraction [Hoffmann et al ICAPS'06b], heuristic 263 functions in model checking [Kupferschmid et al SPIN'06, Hoffmann et al, MoChArt'06], problem decomposition [Köhler&Hoffmann JAIR'00, Porteous et al ECP'01, Hoffmann et al JAIR'04], and creating realistic benchmarks for planning [Hoffmann&Edelkamp JAIR'05, Hoffmann et al JAIR'06] In the context of the KnowledgeWeb WSC prototype deliverable due end Jun'06, Conformant-FF is used as the functional level composition component Since WSMO functional level web service descriptions, like planning actions, are also composed of "preconditions" and "effects", plus "assumptions" and "postconditions", a high-level correspondence is obvious There are, however, large gaps between the respective formalisms behind these keywords: most strikingly, in planning there is no notion of "inputs" and "outputs", and thus, naturally, no notion of ontology at all Instead, as said above planning is formalized, based on the notion of state variables changing their values Since this is a quite fundamental difference, it is our point of view that state-based techniques will be useful as a source of inspiration for WSC, but will not serve to replace targeted methods for WSC One can, however, identify a correspondence between initial state uncertainty and "instance typing" to implement a first tool addressing a very simple form of ontologies, where the concept hierarchy has only two levels, and each "high-level" concept instance will be a member of exactly one "low-level" concept In a nutshell, this is translated into initial state uncertainty by specifying, for each input x, of high-level concept y, of a web service, a clause saying (y_1(x) or or y_k(x)) where y_1 y_k are the sub-concepts of y By doing so, we express that the precise form of the inputs is unknown and may be any of the given list, and that the composed web service should work for all possibilities Conformant-FF then finds a plan corresponding to such a web service (taking the form of a sequence of atomic services) This framework is already extended to "composite" concepts containing other concepts ("name" etc) as their fields In working step of the working group, a prototype shall be implemented, probably exploiting some of FF's techniques The input to the prototype will be a probably in certain ways restricted ontology, a set of WSMO web services in functional level description, and a certain kind of target formula like the existence of an instance of a given concept, satisfying various constraints The output will be a composed service, taking the form of a sequence of the input web services, satisfying the target formula By moving the development out of the planning (state variable based) context and into the web service (ontology based) context, we expect to be able to deal with more complex ontologies than the KnowledgeWeb prototype sketched above In working step of the working group, DERI's project partners, particularly those involved in the SUPER project, shall be contacted to 264 Future Steps try to come up with first ideas on where WSC could be usefully applied A few first WSMO models are expected within half a year In working step 3, the WSC prototype shall be integrated with WSMX, in a clean re-implementation tailored for efficiency and usability In working step 4, first steps shall be taken to design process level composition technology, possibly based on (an extension of) the process level composition part of the KnowledgeWeb prototype, and on the outcome of step In the future, we expect the Composition group's work to revolve around the following topics: Language subset/capability extensions We intend to deal with as large as possible subsets of WSMO/WSML Naturally, the developed technology will start with restricted language subsets, and will incrementally move on to richer subsets This is a process that may well take several years If new features/scenarios become relevant on the side of WSMO/WSML, these will become new targets for WSC This sort of work is, in its nature, very similar to the development the FF system has undergone between 1999 and 2006, moving to ever more complex input languages [Hoffmann&Nebel JAIR'01, Hoffmann ECAI'02, Hoffmann JAIR'03, Brafman&Hoffmann ICAPS'04, Hoffmann&Brafman ICAPS'05, Hoffmann&Brafman AI'06, Domshlak&Hoffmann ICAPS'06] Applications, case studies, benchmarking, technology export A vital ingredient to WSC research will be to stay as close as possible to the envisioned fields of commercial/industrial application We intend to use and strengthen DERI's contacts in this respect Possible areas of application shall be identified, and increasingly realistic scenarios shall be modelled These models play a crucial role in evaluating the developed WSC techniques, and thus guiding the research into which kinds of methods will work and which will not The case studies may eventually lead to fostered collaborations and, ultimately, to technology export Work in this direction modelling applications and benchmarking systems has been done by Jörg Hoffmann in the context of the international planning competition 2004 [Hoffmann&Edelkamp JAIR'05, Hoffmann et al JAIR'06] Addressing efficiency problems Since WSC is a notoriously hard problem even AI Planning in very simple formalisms is PSPACEcomplete it is essential to develop heuristic techniques that have the potential to scale satisfyingly in practical instances of the WSC problem From Jörg Hoffmann's previous work, there is a wide range of techniques to draw upon, ranging from the techniques used in FF, over abstraction techniques [Hoffmann et al ICAPS'06b], and heuristic functions developed in the context of model checking [Kupferschmid et 265 al SPIN'06, Hoffmann et al, MoChArt'06], to problem decomposition methods [Köhler&Hoffmann JAIR'00, Porteous et al ECP'99, Hoffmann et al JAIR'04] Investigations of problem structure, like [Hoffmann AIPS'02, Hoffmann JAIR'05, Hoffmann et al ICAPS'06a], can give, insights on what characterizes the instances where the search techniques work well, ultimately enabling the WSC tool to automatically configure itself in a suitable way Publications We expect that, eventually, notions of optimality will become relevant for WSC: What is the "best" service satisfying the composition task, and how can we compose that service? We intend to contribute to both the development of such notions and to their algorithmic treatment J Köhler and J Hoffmann, On Reasonable and Forced Goal Orderings and their Use in an Agenda-Driven Planning Algorithm, Journal of Artificial Intelligence Research, 12: 338 386, 2000 J Hoffmann and B Nebel, The FF Planning System: Fast Plan Generation Through Heuristic Search, Journal of Artificial Intelligence Research, 14: 253 302, 2001 J Porteous, L Sebastia, and J Hoffmann, On the Extraction, Ordering, and Usage of Landmarks in Planning, Proceedings of the 6th European Conference on Planning (ECP'01), Toledo, Spain, September 2001 J Hoffmann, Local Search Topology in Planning Benchmarks: A Theoretical Analysis, Proceedings of the 6th International Conference on Artificial Intelligence Planning and Scheduling (AIPS'02), Toulouse, France, April 2002 J Hoffmann, Extending FF to Numerical State Variables, Proceedings of the 15th European Conference on Artificial Intelligence (ECAI'02), Lyon, France, July 2002 J Hoffmann, The Metric-FF Planning System: Translating ``Ignoring Delete Lists'' to Numeric State Variables, Journal of Artificial Intelligence Research, 20: 291 341, 2003 J Hoffmann, J Porteous, and L Sebastia, Ordered Landmarks in Planning, Journal of Artificial Intelligence Research, 22: 215 278, 2004 R Brafman and J Hoffmann, Conformant Planning via Heuristic Forward Search: A New Approach, Proceedings of the 14th International Conference on Automated Planning and Scheduling (ICAPS'04), Whistler, Canada, June 2004 266 J Hoffmann, Where Ignoring Delete Lists Works: Local Search Topology in Planning Benchmarks, Journal of Artificial Intelligence Research, 24: 685 758, 2005 J Hoffmann and S Edelkamp, The Deterministic Part of IPC-4: An Overview, Journal of Artificial Intelligence Research, 24: 519 579, 2005 J Hoffmann and R Brafman, Contingent Planning via Heuristic Forward Search with Implicit Belief States, Proceedings of the 15th International Conference on Automated Planning and Scheduling (ICAPS'05), Monterey, CA, USA, June 2005 J Hoffmann and R Brafman, Conformant Planning via Heuristic Forward Search: A New Approach, Artificial Intelligence, 170 (6-7), pages 507 - 541, 2006 J Hoffmann, S Edelkamp, S Thiebaux, R Englert, F Liporace, and S Trueg, Engineering Benchmarks for Planning: the Domains Used in the Deterministic Part of IPC-4, accepted for: Journal of Artificial Intelligence Research C Domshlak and J Hoffmann, Fast Probabilistic Planning Through Weighted Model Counting, accepted for: 16th International Conference on Automated Planning and Scheduling (ICAPS'06) J Hoffmann, C Gomes, and B Selman, Structure and Problem Hardness: Goal Asymmetry and DPLL Proofs in SAT-based Planning, accepted for: 16th International Conference on Automated Planning and Scheduling (ICAPS'06) J Hoffmann, A Sabharwal, and C Domshlak, Friends or Foes? An AI Planning Perspective on Abstraction and Search, accepted for: 16th International Conference on Automated Planning and Scheduling (ICAPS'06) J Hoffmann, J Smaus, A Rybalchenko, S Kupferschmid, and A Podelski, Using Predicate Abstraction to Generate Heuristic Functions in UPPAAL, accepted for: Post-Proceedings of 4th Workshop on Model Checking and Artificial Intelligence (MoChArt'06) S Kupferschmid, J Hoffmann, H Dierks, and G Behrmann, Adapting an AI Planning Heuristic for Directed Model Checking, 13th International SPIN Workshop on Model Checking of Software (SPIN'06) 267 Software releases FFv2.3: Basic version of sb.mpg.de/~hoffmann/ff.html FF, available at: http://www.mpi- Metric-FF: Version of FF dealing with numeric state variables, available at: http://www.mpi-sb.mpg.de/~hoffmann/metric-ff.html Conformant-FF: Version of FF dealing with initial state uncertainty, available at: http://www.mpi-sb.mpg.de/~hoffmann/cff.html Contingent-FF: Version of FF dealing with initial state uncertainty and observation actions, available at: http://www.mpisb.mpg.de/~hoffmann/cff.html Not yet assigned researchers: Name Entry date Cluster Objective Projects Research topic Joerg Hoffmann September 2006 SWS Composition I will be working on the automatic composition of web services from smaller components Based on the description languages defined by WSML, services will be composed both on the functional level and on the process level To accomplish this, I will draw on methods from AI, precisely from Automated Planning, in which field I have been working since 1999 Key methods for functional level composition will be the definition of language hierarchies, the identification of maximal tractable fragments, and heuristic search Regarding process level composition, I will work in close collaboration with ITC-IRST, Trento, Italy, who have already established a set of preliminary tools for this purpose, based on Binary Decision Diagrams Since I completed my PhD in 2002, I have published articles in Progress JAIR and AI, the two leading top-quality journals in Artificial towards Habil Intelligence Further articles are pending I plan to complete a Habil within the next years Implementations I have implemented a large number of systems for Automated Planning, and for Model Checking The most important ones are: IPP (1997-1999) Based on heuristic search Returns optimal plans for a propositional planning language Won a 1st prize at the 1998 international planning competition (IPC) FF (1999-2000) Based on heuristic search Treats a propositional planning language Brought about a major breakthrough in scalability of planning systems, in particular outperforming all other systems at the 2000 IPC 268 Publications Metric-FF (2001) Extension to handle numeric language constructs Top performer at 2002 IPC Conformant-FF, Contingent-FF (2003-2005) Extensions to handle uncertainty and observations Best paper award for Conformant-FF at ICAPS 2004, top performer in 2006 IPC Probabilistic-FF (2006) Extension to handle probability distributions Heuristic UPPAAL (2004-2006) Modification of the state-ofthe-art Model Checker UPPAAL to use automatically generated heuristic functions Two versions implemented, one based on ideas from FF, another based on an abstraction method from Model Checking SATPLAN (2004-2006) Based on satisfiability testing Returns optimal plans for a propositional planning language Won the tracks for optimal planners at the 2004 and 2006 IPCs Book (peer-reviewed) J Hoffmann: Utilizing Problem Structure in Planning: A Local Search Approach, LNCS 2854, Springer-Verlag, 2003 Journal Papers (all peer-reviewed) J Hoffmann, S Edelkamp, S Thiebaux, R Englert, F Liporace, and S Trüg: Engineering Benchmarks for Planning: the Domains Used in the Deterministic Part of IPC-4, in: Journal of Artificial Intelligence Research (JAIR), Volume 26, 2006, pages 453 - 541 J Hoffmann and R Brafman: Conformant Planning via Heuristic Forward Search: A New Approach, in: Artificial Intelligence (AI), Volume 170 (6-7), 2006, pages 507 - 541 J Hoffmann: Where Ignoring Delete Lists Works: Local Search Topology in Planning Benchmarks, in: Journal of Artificial Intelligence Research (JAIR), Volume 24, 2005, pages 685 – 758 J Hoffmann and S Edelkamp: The Deterministic Part of IPC-4: An Overview, in: Journal of Artificial Intelligence Research (JAIR), Volume 24, 2005, pages 519 – 579 S Thiebaux, J Hoffmann, and B Nebel: In Defense of PDDL Axioms, in: Artificial Intelligence (AI), Volume 168 (1-2), 2005, pages 38 - 69 269 J Hoffmann, J Porteous, L Sebastia: Ordered Landmarks in Planning, in: Journal of Artificial Intelligence Research (JAIR), Volume 22, 2004, pages 215 – 278 J Hoffmann: The Metric-FF Planning System: Translating ``Ignoring Delete Lists'' to Numeric State Variables, in: Journal of Artificial Intelligence Research (JAIR), Volume 20, 2003, pages 291 - 341 J Hoffmann and B Nebel: The FF Planning System: Fast Plan Generation Through Heuristic Search, in: Journal of Artificial Intelligence Research (JAIR), Volume 14, 2001, pages 253 - 302 Recipient of the 2005 IJCAII-JAIR best paper prize J Rintanen and J Hoffmann: An Overview of Recent Algorithms for AI Planning, in: Künstliche Intelligenz, Volume 2/01, 2001, pages – 11 J Koehler and J Hoffmann: On Reasonable and Forced Goal Orderings and their Use in an Agenda-Driven Planning Algorithm, in: Journal of Artificial Intelligence Research (JAIR), Volume 12, 2000, pages 338 – 386 Conference Papers (all peer-reviewed) J Hoffmann, J Smaus, A Rybalchenko, S Kupferschmid, and A Podelski: Using Predicate Abstraction to Generate Heuristic Functions in UPPAAL, in: Post-Proceedings of the 4th Workshop on Model Checking and Artificial Intelligence (MoChArt'06) J Hoffmann, A Sabharwal, and C Domshlak: Friends or Foes? An AI Planning Perspective on Abstraction and Search, in: Proceedings of the 16th International Conference on Automated Planning and Scheduling (ICAPS'06) J Hoffmann, C Gomes, and B Selman: Structure and Problem Hardness: Goal Asymmetry and DPLL Proofs in SAT-based Planning, in: Proceedings of the 16th International Conference on Automated Planning and Scheduling (ICAPS'06) C Domshlak and J Hoffmann: Fast Probabilistic Planning Through Weighted Model Counting, in: Proceedings of the 16th International Conference on Automated Planning and Scheduling (ICAPS'06) 270 S Kupferschmid, J Hoffmann, H Dierks, and G Behrmann: Adapting an AI Planning Heuristic for Directed Model Checking, in: Proceedings of the 13th International SPIN Workshop on Model Checking of Software (SPIN'06) J Hoffmann and S Kupferschmid: A Covering Problem for Hypercubes, in: Poster Proceedings of the 19th International Joint Conference on Artificial Intelligence (IJCAI'05) J Hoffmann and R Brafman: Contingent Planning via Heuristic Forward Search with Implicit Belief States, in: Proceedings of the 15th International Conference on Automated Planning and Scheduling (ICAPS'05) S Trueg, J Hoffmann, and B Nebel: Applying Automatic Planning Techniques to Airport Ground-Traffic Control: A Feasibility Study, in: Proceedings of the 27th German Conference on Artificial Intelligence (KI'04) R Brafman and J Hoffmann: Conformant Planning via Heuristic Forward Search: A New Approach, in: Proceedings of the 14th International Conference on Automated Planning and Scheduling (ICAPS'04) Recipient of the best paper award S Thiebaux, J Hoffmann, and B Nebel: In Defense of PDDL Axioms, in: Proceedings of the 18th International Joint Conference on Artificial Intelligence (IJCAI'03) J Hoffmann and H Geffner: Branching Matters: Alternative Branching in Graphplan, in: Proceedings of the 13th International Conference on Automated Planning and Scheduling (ICAPS'03) J Hoffmann: Extending FF to Numerical State Variables, in: Proceedings of the 15th European Conference on Artificial Intelligence (ECAI'02) J Hoffmann: Local Search Topology in Planning Benchmarks: A Theoretical Analysis, in: Proceedings of the 6th International Conference on Artificial Intelligence Planning and Scheduling (AIPS'02) J Hoffmann and B Nebel: RIFO Revisited: Detecting Relaxed Irrelevance, in: Proceedings of the 6th European Conference on Planning (ECP'01) 271 J Posteous, L Sebastia, and J Hoffmann: On the Extraction, Ordering, and Usage of Landmarks in Planning, in: Proceedings of the 6th European Conference on Planning (ECP'01) J Hoffmann: Local Search Topology in Planning Benchmarks: An Empirical Analysis, in: Proceedings of the 17th International Joint Conference on Artificial Intelligence (IJCAI'01) J Hoffmann, A Heuristic for Domain Independent Planning and its Use in an Enforced Hill-climbing Algorithm, in: Proceedings of the 12th International Symposium on Methodologies for Intelligent Systems (ISMIS'00) J Hoffmann and J Koehler: A new Method to Index and Query Sets, in: Proceedings of the 16th International Joint Conference on Artificial Intelligence (IJCAI'99) J Koehler, B Nebel, J Hoffmann, and Y Dimopolous: Extending Planning Graphs to an ADL Subset, in: Proceedings of the 4th European Conference on Planning (ECP'97) Workshop Papers (all peer-reviewed) J Hoffmann, C Gomes, and B Selman, Structure and Problem Hardness: Asymmetry and DPLL Proofs in SAT-Based Planning, in: Proceedings of the Workshop on Constraint Propagation and Implementation at CP'05 B Becker, M Behle, F Eisenbrand, M Fraenzle, M Herbstritt, C Herde, J Hoffmann, D Kroening, B Nebel, I Polian, and R Wimmer: Bounded Model Checking and Inductive Verification of Hybrid Discrete-continuous Systems, in: Proceedings of the Workshop ''Methoden und Beschreibungssprachen zur Modellierung und Verifikation von Schaltungen und Systemen'' R Brafman and J Hoffmann: Conformant Planning via Heuristic Forward Search, in: Proceedings of the Workshop on Planning under Uncertainty and Incomplete Information at ICAPS'03 S Edelkamp and J Hoffmann: Quo Vadis, IPC-4? Proposals for the Classical Part of the 4th International Planning Competition, in: Proceedings of the Workshop on the Competition at ICAPS'03 272 S Thiebaux, J Hoffmann, and B Nebel: In Defense of PDDL Axioms, in: Proceedings of the Workshop on the Competition at ICAPS'03 J Hoffmann and B Nebel: What Makes the Difference Between HSP and FF?, in: Proceedings of the Workshop on Empirical Methods in AI at IJCAI'01 J Hoffmann and B Nebel: Towards Thorough Empirical Methods for AI Planning, in: Proceedings of the Workshop on Empirical Methods in AI at IJCAI'01 J Koehler and J Hoffmann: On the Instantiation of ADL Operators Involving Arbitrary First-Order Formulas, in: Proceedings of the Workshop on New Results in Planning, Scheduling and Design (PuK2000) at ECAI 2000 J Hoffmann: A Heuristic for Domain Independent Planning and its Use in an Enforced Hill-climbing Algorithm, in: Proceedings of the Workshop on New Results in Planning, Scheduling and Design (PuK2000) at ECAI 2000 J Koehler and J Hoffmann: Planning with Goal Agendas, in: Proceedings des 13 Workshop Planen und Konfigurieren auf der 10 Tagung Expertensysteme (XPS-99) 273