EvoSTIM2003
5th European Workshop on Scheduling and Timetabling

The 4 papers accepted for last year's EvoSTIM workshop are available in Springer's LNCS series, volume 2279.

Previous editions:
Göteborg, Sweden, 1999
Edinburgh, UK, 2000
Lake Como, Italy, 2001
Kinsale, Ireland, 2002

Introduction

EvoSTIM2003 is the fifth workshop of the EvoNet working group on scheduling and time tabling.

Scheduling and timetabling are amongst the most successful applications of evolutionary techniques. A related and promising field for the application of evolutionary methods is AI planning.

EvoSTIM2003 represents a unique opportunity for evolutionary computing researchers in the inter-related areas of planning, scheduling and timetabling to come together, present their latest research and discuss current developments and applications.

The workshop proceedings will be published by Springer in the LNCS series and will be available at the workshop.

A special issue of the Journal of Mathematical Modelling and Algorithms on Evolutionary Computation in Combinatorial Optimization (see http://www.ads.tuwien.ac.at/ecco_jmma) will be published. Authors of selected EvoSTIM papers will be invited to submit an extended version for consideration in this special issue.

Topics of interest

Topics of interest include, but are not limited to:

• case studies of applications of evolutionary techniques to real-life problems in scheduling, timetabling, and AI planning
• theoretical developments, including benchmarking and other comparative studies
• hybrid methods and novel heuristics
• methods for fast and minimally disruptive rescheduling
• dynamic scheduling
• distributed scheduling
• scheduling in industrial environments including transport and distribution
• educational courses and exam timetabling
• employee and workforce scheduling
• organisation and resource scheduling
• generation of plans and construction of planners
• plan merging and decomposition of planning problems
• handling multiple goals in scheduling and planning
• planning under resource constraints.

Draft: subject to change

See also: Programme overview

The EvoWorkshops2003 proceedings will be
published by Spinger as part of their
Lecture Notes in Computer Science series.

DNA based algorithms for some scheduling problems
Blazewicz J, Formanowicz P, Urbaniak R
RNA computing is an alternative approach to performing computations. In general it is possible to design a series of biochemical experiments involving DNA molecules which is equivalent to making transformations of information. In classical computing devices electronic logical gates are elements which allow for storing and transforming information. Designing of an appropriate sequence or a net of ''store'' and ''transform'' operations (in a sense of building a device or writing a program) is equivalent to preparing some computations. In DNA computing the situation is analogous, and the main difference is that instead of electronic gates DNA molecules are used for storing and transforming information. From this follows that the set of basic operations is different in comparison to electronic devices but the results of using them may be similar. In this paper DNA based algorithms for solving some single machine with limited availability scheduling problems are presented. To our best knowledge it is the first attempt to solve scheduling problems by molecular algorithms.
EvoSTIM Session 1: : April 14, 1130-1300

Learning Action Strategies for Planning Domains using Genetic Programming
Levine J, Humphreys D
There are many different approaches to solving planning problems, one of which is the use of domain specific control knowledge to help guide a domain independent search algorithm. This paper presents L2Plan which represents this control knowledge as an ordered set of control rules, called a {\it policy}, and learns using genetic programming. The genetic program's crossover and mutation operators are augmented by a simple local search. L2Plan was tested on both the blocks world and briefcase domains. In both domains, L2Plan was able to produce policies that solved all the test problems and which outperformed the hand-coded policies written by the authors.
EvoSTIM Session 1: : April 14, 1130-1300

Routing using Evolutionary Agents and Proactive Transactions
Urquhart N, Ross P, Paechter B, Chisholm K
The authors have previously introduced the concept of building a delivery networ k using an agent-based system. The delivery networks are built in response to a real-world problem that involves delivering post to a large number of households within an urban area. The initial agent based system worked to primarily resolv e hard constraint violations. To further improve the solution obtained by the ag ents, we propose to allow agents to negotiate exchanges of work. We demonstrate the solution obtained may be further improved by allowing such negotiated transa ctions.
EvoSTIM Session 1: : April 14, 1130-1300

Chair

Emma Hart <e.hart@napier.ac.uk>
Martin Middendorf <mmi@aifb.uni-karlsruhe.de>

Programme committee

• Andrea Tettamanzi (Italy)
• Daniel Merkle (Germany)
• Daniel Borrajo (Spain)
• Peter Ross (UK)
• Ben Paechter (UK)vJohn Levine (UK)
• Peter Swann (UK)