IT-DepartmentÅbo Akademi

UniversityPresentation for Academy of

Finland evaluation

Barbro BackDepartment chair

IT in Turku

University of Turku /Math

Turku School of Economics/

IS

Åbo AkademiUniversity / IT

University of Turku / IT

Turku Polytechnic /IT

TUCSICT-house

IT-department

New Faculty of Technology at Åbo Akademi 2006 --

Two departments IT (Information technology) KT (Chemical engineering)

IT previously spread out over three different faculties MNF, KTF, ESF

Present personnel Teaching appr. 30

8 professors, 10 Lecturers,11 others Administration, technical support 9 Research 60 – 80 persons

IT-department (cont.)

Students and Degrees Approx. 600 (masters and Ph.D. students) 60 - 70 new students and 10-12 new doctoral students enrolled each year

Financial volume 2006 Internal funding : 2 Meuros External funding : 1.8 Meuros

Output 164 publications in 2006

30 in refereed journals 77 in refereed conference publications

8 Doctoral theses in 2006 53 MSc theses in 2006

IT at Åbo Akademi 2006-

Åbo Akademi

Faculty of Technology

IT department

Department of Chemical Engineering

CREST

IAMSR

Researchgroups

45

35

13020

CREST

Center for Reliable Software Technology Leader: Academy professor Ralph-Johan Back Focus: how to model, construct and analyze large reliable

software intensive systems Distributed systems lab: Kaisa Sere + 15

researchers Embedded systems lab: Johan Lilius + 15 Software construction lab: Ralph Back, Ivan

Porres + 10 Learning and reasoning lab: Ralph Back, Tapio

Salakoski (UTU)+ 5 Gaudi software factory: Ralph Back, Ivan Porres

(in summer 20 programmers)

IAMSR Institute for Advanced Management Studies

Leader:professor Christer Carlsson Focus: Combines topics from management science and

information systems Soft Computing: Christer Carlsson + 25

fuzzy logic, fuzzy optimization, real options and fuzzy real options modeling

intelligent MAS for industry foresight risk assessment in grid computing advanced logistics planning and optimisation

Mobile technologies applications: Pirkko Walden + 10 knowledge mobilization and mobile value services and

applications IAMSR was on shortlist for CoE 2008 - 2013

Other research labs and groups

Computational Biomodeling lab: Ion Petre + 6 researchers

Data Mining and Knowledge Management lab: Barbro Back + 6

High-Performance Computing lab: Jan Westerholm + 4

Process Control lab: Hannu Toivonen + 4

Ph.D. theses 2000-2007 Linas Laibinis (J. von Wright) Martin Büchi (R.Back) Elena Troubitsyna (K.Sere) Shuhua Liu (C.Carlsson) Ivan Porres (R.Back) Mauno Rönkkö (K.Sere) Rainer Jansson (C.Carlsson) Vladimir Kvassov (C.Carlsson) Franck Tétard (C.Carlsson) Bill Anckar (P. Walden) Antonina Kloptchenko (B.Back) Eija Koskivaara, TSE (B.Back) Péter Majlender (C.Carlsson) Otieno Mbare (C.Carlsson) Mikael Collan (C.Carlsson) Tomas Eklund (B.Back) Aapo Länsiluoto, TSE (B.Back) Rimvydas Ruksenas (K.Sere)

Seppo Wirtanen, UTU (J.Lilius) Dag Björklund (J.Lilius) Cristina Seceleanu (R.Back) Luigia Petre ((K.Sere, R.Back) Lu Yan (K.Sere) Shengnan Han (P. Walden) Jukka Arvo, UTU, (J. Westerholm) Irina Georgescu (C.Carlsson) Adrian Costea (B.Back) Orieta Celiku (J. von Wright) Viorel Preoteasa (R.Back) Luka Milanov (R.Back) Ville Harkke (P.Walden) Chihab BenMoussa (C.Carlsson) Kaj-Mikael Björk (C.Carlsson) Francisco Augusto Alcaraz Garcia

(C.Carlsson) Erkki Patokorpi (P. Walden) Dragos Truscan (J.Lilius) Markus Alanen (I.Porres)

Research highlights

Short presentation of main research activities

Focus on Computer Science and Engineering Can only present a few research highlights

Many excellent projects going on Answer questions after presenting research highlights

CREST backgroundRalph Back

Refinement calculus 1978 – (Ralph Back) Action systems 1982 – (Ralph Back, Reino Kurki-

Suonio) Programming methods group at Åbo Akademi

1984 - Refinement calculus development (Back, von Wright) Action System development (Back, Sere)

PMG splits up into five research labs 2002 Four labs form Crest High performance computing lab, Westerholm

Crest Center of Excellence in Research 2002 – 2007

Crest was on shortlist for CoE 2008 - 2013 Reviews were excellent, but competition very strong

Teaching mathematics in high school

Ralph Back Tapio Salakoski

Structured derivations in high school

Structured derivations: Based on Dijkstra’s calculational style for proofs Extended to complete logical system based on sequent calculus

(Back and von Wright: Refinement calculus, Springer 1998) Validated in Kupittaa high school pilot study

3 year study (test group using structured derivations, control group using standard approach). All courses in high school curriculum covered

Test group clearly and consistently outperformed control group Pilot projects in a number of other high schools schools

Supported by large course development and teacher training grant New book: Mathematics with a little bit of logic

Back and von Wright Presents basic approach as a high school course, lots of examples

Structured derivations exampleSolve equation |x-1|+|2x-y|=0

|x-1|+|2x-y|=0 {property of absolute

values}x-1=0 2x-y=0

{add 1 to both sides of left equation}x=1 2x-y=0

{simplify right conjunct}• [x=1]

2x-y=0 {substitute x=1 by

assumption}2-y=0

{solve}y=2

• x=1 y=2

motivate each step, on separate line

nested derivation

logical notation in derivation

assumptions have scope

collapsabel outlines

Publish on web

Invariant based programming

Ralph Back

Invariant based programming Constructing programs and correctness proofs

together Based on early work by Back and early TR by John Reynolds Start by formulating the invariants of the system, then

transitions between invariants, and then verify correctness of transitions

Formal methods light Easy to learn, easy to use in practice Have carried out a number of sessions teaching both novices and

experienced programmers how to construct verified programs Now teaching approach to first year CS students at Åbo Akademi

Socos environment provides computer support for method Drawing diagrams, proving verification conditions automatically,

generating executable code, testing and debugging

Nested invariant diagrams

A: array 1..n of integer; Permutation(A,A0); 1 <= n

[A[j] >= A[k]

i: integer; 1 <= k <= n; Sorted(A,1,i-1),Partitioned(A,i)

[i=/=n] j,k:=1,1j,k: integer; A[k]=min(A,i,j);k < n; 1 <= k <= j <= n

[j=/=n]

j:=j+1

[A[j] < A[k]; k:= j

Sorted(A,1,n)

[j=n]i:=i+1A[i],A[k]:=A[k],A[i]

[i= n]

i:= 1

Sort array A

Organize program in nested invariants

Transitions between invariants: [guard] assignments

Formulate invariants before adding transitions

Verify correctness of program as it isbeing constructed

High degree of locality in constructing,modifying and verifying program

Gaudi Software factory

Ralph Back Ivan Porres

Software Experience and Software Development Factory

Development Factory

Development Factory

Experience Factory

Experience Factory

ResearchLaboratories

ResearchLaboratories

IndustryIndustry

EducationEducation

Develop tools as demonstrators

Software Process ImprovementSandbox

Improve courses on Software Engineering

Research focus

Industrial projects Software Process Improvement Efforts

Medium-term projects Software Development and Experience Factory Setting Agile and team development processes

Long-term research Incremental software development & software reuse Software architectures Language mechanisms Domain-specific modeling languages

Distributed Systems Design

Kaisa Sere, Marina Walden, Elena Troubitsyna

System approach - from theory to practice

System approach A system or a computer-based product is first modeled as one

entity on a very abstract level. Thereafter, detail is introduced into the model in a systematic

manner (e.g. using patterns). When the model is detailed enough it is decomposed into

meaningful parts or sub-models, e.g. environment, hardware and software.

Dependability: correctness (fault avoidance) , fault tolerance

Formalisms: Action systems, B Method, Event B The approach gives structure to the design

of complex distributed systems and helps in managing the derivation task.

Long-term financing via EU projects

MATISSE (IST FP5 2000-03) Methodologies and technologies for industrial strength

systems engineering developed a design methodology for critical control systems

using formal methods RODIN (IST FP6 STREP 2004-07)

Rigorous open development environment for complex systems creation of a methodology and a supporting open tool platform

for the cost effective development of dependable complex software systems and services

DEPLOY (IST FP7 IP 2008-12) : Industrial deployment of advanced system

engineering methods for high productivity and dependability

Status: hearings in Brussels tomorrow

System approach collaborators

ETH/Z (Ch) Jean-Raymond Abrial, David Basin

Univ. of Newcastle (UK) Cliff Jones, Alexander Romanovsky

Univ. of Southampton (UK) Michael Butler, Michael Leuschel

Industrial Nokia (Fi), Siemens (Fr), PerkinElmer (US/Fi), Atec (UK),

Praxis (UK), Gemplus (Fr), Qinetiq (UK), ClearSy (Fr)

System approach results 2 PhD theses 1995 - 1999

Marina Waldén: Formal reasoning about distributed systems

Virpi Kasurinen: Informal and formal requirements specification

Program Development by Stepwise Refinement (E. Sekerinski, K. Sere (Eds.), Springer 1998)

4 PhD theses 2000 - 2006 Elena Troubitsyna: Stepwise development of

dependable systems Mauno Rönkkö: Stepwise development of hybrid

systems Lu Yan: Systematic design of ubiquitous systems Luigia Petre: Modelling with action systems

2 PhD theses will be evaluated summer 2007

Pontus Boström, Dubravka Ilic

Model Driven Development

Johan Lilius Ivan Porres

Model-Driven Software and System Development

TheoryTheory

Tool FrameworksTool Frameworks

ApplicationsApplications

The goal of the Model Driven Development group at CREST is to study and develop new ways to build large software-intensive systems in practice using advanced modeling languages and tools

MDD

Long term research goals: Understand methods for domain driven design,Understand the foundations (structure and algorithms)of domain driven design. Results:

A methodological framework verified through 2 substantial case studies (PhD Dragos Truscan)

A framework for the semantics of DSL that enabled codegeneration (PhD Dag Björklund).

A framework for metamodelling (PhD Marcus Alanen)

MDD (cont)

Middle term research goals: Implementation of theoretical results in tool frameworks Results

CORAL - A tool to define, edit and transform visual modeling languages: Modeling language definition, Visual languages and editors, Model transformation, Model interchange

Rialto - A language for multiple models of computation: Flexible definition of behavioral semantics

MDD(cont.) Short term research goals: Verify our more theoretical

results in practice, Apply the tool frameworks, provide feedback to middleand long term goals Results

MICAS platform: Tool for definition of mobile phone subsystems

PECOS: Tools and methods for verification of non-functional constraints (energy) in mobile phone subsystems

Future topics Model based testing Dealing with non-functional constraints in system design Version control

Computational systems biology

Ion Petre

Gene assembly in ciliates

Models for gene assembly Linked lists in living cells! Invariants Complexity Parallelism Applications in genetic

engineering Collaborators:

Gregorz Rozenberg (CS,Holland)

A.Ehrenfeucht (Math,US) D.M.Prescott (Biol,US) Tero Harju (Math,FIN)

Heat shock response

The cells reaction to elevated temperature

Models Continuous ODE

models Stochastic models

Collaborators: Ralph Back(CS, FIN) John Eriksson(Biol,

FIN) Lea Sistonen(Biol, FIN) Rick Morimoto(Biol,

US)

Heat shock genePromoter

Heat shock genePromoter

HSPHSP

HSPHSP

HSF

RNA pol

HSP:HSF

MFP

MFP

MFPMFP

MFP

37C

42C

Thank You