Grid Enabled Optimisation and Design Search for Engineering (GEODISE)

1
Grid Enabled Optimisation and Design Search for
Engineering (GEODISE)

• http//www.geodise.org

2
Academic and Industrial PartnersSouthampton,
Oxford and Manchester

• Simon Cox- Grid/ W3C Technologies and High
  Performance Computing
• Global Grid Forum Apps Working Group
• Andy Keane- Director of Rolls Royce/ BAE Systems
  University Technology Partnership in Design
  Search and Optimisation
• Mike Giles- Director of Rolls Royce University
  Technology Centre for Computational Fluid
  Dynamics
• Carole Goble- Ontologies and DARPA Agent Markup
  Language (DAML) / Ontology Inference Language
  (OIL)
• Nigel Shadbolt- Director of Advanced Knowledge
  Technologies (AKT) IRC

• BAE Systems- Engineering
• Rolls-Royce- Engineering
• Fluent- Computational Fluid Dynamics
• Microsoft- Software/ Web Services
• Intel- Hardware
• Compusys- Systems Integration
• Epistemics- Knowledge Technologies
• Condor- Grid Middleware

http//www.geodise.org
3
Geodise-K
Geodise will provide grid-based seamless access
to an intelligent knowledge repository, a
state-of-the-art collection of optimisation and
search tools, industrial strength analysis codes,
and distributed computing data resources
4
Design
5
Design Challenges

• Modern engineering firms are global and
  distributed
• Not just a problem of using HPC

How to ?
CAD and analysis tools, user interfaces, PSEs,
and Visualization
improve design environments cope with legacy
code / systems
Optimisation methods
produce optimized designs
Management of distributed compute and data
resources
integrate large-scale systems in a
flexible way
Data archives (e.g. design/ system usage)
archive and re-use design history
Knowledge repositories knowledge capture and
reuse tools.
capture and re-use knowledge
6
Design and the Grid
PSE on Steroids
7
KNOWLEDGE
INFORMATION
DATA
COMPUTATION
8
Design of Experiment Response Surface Modelling
9
NASA Satellite Structure

• Optimized satellite designs have been found with
  enhanced vibration isolation performance using
  parallel GAs running on work-station clusters.

10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
Baseline 3D-boom on test
16
Design Optimisation Methods

• OPTIONS is
• a design exploration and optimization package
  that may be used to study and compare a large
  range of optimization methods when applied to
  design problems.
• Has gt 30 algorithms e.g.
• Classical gradient descent methods.
• Evolutionary and stochastic searches.
• Response surface models.
• Data-fusion methods.
• Multi-objective / Pareto approaches.
• Robust Engineering Design (RED) methods.

17
KNOWLEDGE
INFORMATION
(DATA)
COMPUTATION
18
Automated Data Archiving
19
From information to knowledge

• Information
• Which resources have the codes I require?
• Give me the current status of all resources
• When are resources under-used?
• What is the status of my jobs?
• Where was the code when the machine crashed?
• When will I get my results?

• Re-use and analysis of knowledge
• Why do these parameters perform well?
• What similar designs have been studied before?
• Which systems give good performance on which
  jobs?
• How to use the package?
• Who is an expert user of Genetic Algorithms?
• Which design strategies are likely to prove
  effective?

20
KNOWLEDGE
INFORMATION
DATA
COMPUTATION
21
Knowledge repositories.

• Rule bases on available methods
• Case based reasoning systems
• Ontologies for allow use of domain and process
  knowledge
• Search and retrieval systems
• Intelligent resource management

22
Average Normalized Evaluation Count
Best-Quality Optimizer, SA
Average Normalized Objective Function Value
Pictorial representations of the Overall
Performance of each Search Technique on the
Aircraft Wing Design Case Study
23
Geodise-W
24
Geodise-K
Geodise will provide grid-based seamless access
to an intelligent knowledge repository, a
state-of-the-art collection of optimisation and
search tools, industrial strength analysis codes,
and distributed computing data resources
25
Technologies (i) Grid Middleware

• (To coordinate and authenticate use of components
  of Geodise)
• Globus (and GGF grid-computing protocols)
• Security Infrastructure (GSI)
• Resource Allocation Mechanism (GRAM)
• Resource Information System (GRIS)
• Index Information Service (GIIS)
• Grid-FTP
• Metadirectory service (MDS 2.0) coupled to LDAP
  server
• Condor (distributed high performance throughput
  system)
• Condor-G allows us to handle dispatching jobs to
  our Globus system
• Active collaboration from with the Condor
  development team at University of Wisconsin
  (Miron Livny)

26
(ii) Data Open W3C Standards

• (To access and interchange data)
• XML and XML Schema
• Representing data in a portable format
• WSDL (Web Service Description Language)
• UDDI (Universal Description, Discovery and
  Integration)
• Publish and discover information about web
  services

27
(iii) Ontologies Semantic Web

• (conceptualisation of a communitys knowledge of
  a domain)
• DAML - OIL (DARPA Agent Markup Language/ Ontology
  Inference Language)
• Genetics http//www.geneontology.org/
• Virtual Enterprises
• Product Specifications
• Medicine
• Encyclopaedic Knowledge
• http//www.cyc.com/cyc-2-1/toc.html

28
(iv) Knowledge Technologies
29
The future of design optimisation

• Design improvements driven by the exploitation of
  CAD tools coupled to advanced analysis codes
  (CFD, FEA, etc.)
• Distributed, heterogeneous computing environment
  spread across companies and time zones.
• Optimization used alongside manual search as part
  of a problem solving environment.
• Knowledge based tools for advice and control of
  process as well as product.

30
Conclusions

• Design Optimisation needs all layers of Grid
• Computation
• Data
• Information
• Knowledge
• Geodise will provide grid-based seamless access
  to an intelligent knowledge repository, a
  state-of-the-art collection of optimisation and
  search tools, industrial strength analysis codes,
  and distributed computing and data resources