PLANSERVE%20Planning%20and%20Scheduling%20Techniques%20for%20the%20Intelligent%20Problem%20Solving%20Grid

1
PLANSERVEPlanning and Scheduling Techniques for
the Intelligent Problem Solving Grid

• Planning and Scheduling Team
• ISTC-CNR
• National Research Council of Italy
• Viale Marx 15, I-00137 Rome, Italy
• http//pst.ip.rm.cnr.it

2
Planning and Scheduling _at_ ISTC-CNR

• Different aspects of research and development
• Algorithms for Planning and Scheduling (PS)
• integration of PS algorithms
• Constraint-based reasoning
• temporal and resource reasoning
• Interactive Problem Solving
• how to include the user in the loop
• Space applications
• OOSCAR project (ASI - Italian Space Agency)
• MEXAR Interactive Support for Mission Planning in
  MARS EXPRESS (ESA project ended in July 2002)
• Collaborations
• University of Rome La Sapienza
• Robotics Institute, Carnegie Mellon University,
  Pittsburgh, USA

3
Present Projects

• ASI (Italian Space Agency) -- Basic research
  funds
• SACSO SAfety Critical Software for robotics
  (JERRY)
• Contraint-Based Continuous Planning (O-OSCAR and
  beyond)
• DOVES
• ARISCOM
• MIUR (Italian Ministry for Education University
  and Research)
• Multi-agent systems with software and robotic
  agents for socially relevant applications

4
PST Members

• CNR Research Scientists
• Amedeo Cesta
• Angelo Oddi
• PhD Students
• Gabriella Cortellessa
• Nicola Policella
• Simone Fratini
• Research Assistants
• Federico Pecora
• Riccardo Rasconi

5
IPSG Prototype our role
6
Potential Contributions

• Definition of the class of Planning and
  Scheduling (PS) problems involved.
• A particular case is the integration of planning
  and scheduling algorithms
• Solver Agent
• Representation of the solver configurations
• Domain analysis and mapping problems in to solver
  configurations
• Selection and integration of the software modules
• Software infrastructure

7
Planning and Scheduling Problems

• A first problem is the definition of the class of
  scheduling and planning problems that we can
  solve.
• A selection from the current literature of the
  best planning and scheduling algorithms. This
  selection must be consider the availability of
  free and robust implementation of the algorithms.

8
Planning and Scheduling Solvers

• A particular case is the integration of planning
  and scheduling algorithms.
• We have some experience on component integration
  for scheduling architectures and are doing right
  now some research for understanding different
  ways for integrating a generic planner with one
  or more scheduling algorithms.
• Classical Planners OOSCAR
• Graph Plan-based planners OOSCAR

9
Domain Analysers

• A Domain Analyser should bridge the gap between
  problems and solver configurations. Several
  issues are involved
• Classification of the set of possible PS domain
  problems (e.g, with respect to the type of
  temporal and resource constraints, the level of
  temporal flexibility, the degree of parallelism
  of the actions, etc.)
• A representation of the solver configurations
• A mechanism for mapping problems in to solvers

10
Software Infrastructure

• A fundamental problem is how "put" together all
  the modules in a way that they can communicate
  each other, refer to a common representation of
  the problems and be combined in different ways.
• Several issues are involved in this phase, one is
  the choice of a software infrastructure for the
  communication
• For example, ACE (the Adaptive Communication
  Environment)

11
Comments and Conclusions

• Integration of PS algorithms should have an
  important role
• We still lack of some strong examples
• the role of the PS technologies in the SMEs
• with and without KE services
• HCI techniques should have a greater role
• Advanced features
• mixed-initiative tool for the generation of
  solver configurations
• dynamic and adaptive mapping between problems and
  solvers.