About IAM

Mission

The Industrial Applications and Methods (IAM) Laboratory deploy, analyse, develop, and implement algorithmic methods from a wide selection of fields in computer science. The qualifying characteristic of methods chosen is that they should be suited for real world, large scale industrial applications. Currently this includes methods for planning and scheduling, resource allocation and flow optimisation, data analysis, process modelling and monitoring, fault diagnosis and decision support. This research direction is motivated by the conviction that the science of computation must include the study of practical application of its results. We aim to collect, study, apply, classify, improve and when necessary invent new algorithmic techniques and methods motivated by individual applications. We also aim to widen the field of applications in industry by applying algorithmic methods to a new problem domains within industry.

Research goals

Many of the algorithmic methods studied in individual areas of computer science have been applied only to idealisations of real life industrial problems. To solve or support the solution of real practical industrial problems it is generally necessary to analyse the full problem very carefully and try and understand it in terms of well understood subproblems.

This activity leads to two types of results:

  1. Better understanding of typical models for a selection of important real-life industrial problems
  2. Better understanding of the properties of a selection of practically useful algorithmic methods

This said, the choice of actual industrial applications at each point in time reflects the backgrounds and experiences of the scientists employed by the laboratory. The current state of the art in the laboratory can thus be characterised in two dimensions:

  1. The types of problems of which we have working knowledge
  2. The types of algorithmic methods we use and investigatE

Problems currently investigated

  • Capacity and network analysis and flow optimisation in e.g. transportation, logistics and telecom
  • Classification and diagnosis in in e.g. bioinformatics
  • Decision support in chemical analysis and synthesis with applications in e.g. drug desigN
  • Infrastructure design support with applications in telecom and transportation
  • Fault detection and analysis with applications in e.g. process and industry
  • Matching and structure detection in bio-informatics
  • Analysis and usage optimisation with application in e.g. telecom and transportation
  • Process planning and monitoring for process and manufacturing industry with applications in e.g. steel, paper and petrochemical industries * Resource allocation in transportation and telecom applications
  • Statical and dynamic scheduling for personnel, vehicles and or other production resources in e.g transportation
  • Traffic analysis and optimisation applications in e.g. transportation, telecom and energy distribution
  • Scheduling and flow optimisation for e.g energy production and distribution

Algorithmic methods currently used

  • Bayesian statistics
  • Combinatory reasoning
  • Constraint programing
  • Discrete event systems
  • Information theory
  • Learning systems
  • Local search methods
  • Mathematical logic and algebra
  • Matching theory
  • Operations research
  • Scheduling methods
  • Statistical modelling methods

Typical results

Depending on the type of application and the type of results expected working methods

  • Models of practical industrial processes
  • Functional requirements and prototypes of support systems
  • Solution methods for diagnosis, planning, resource allocation, scheduling and structure analysis
  • Methodology for application of algorithmic methods
  • Prototype implementations of novel solver algorithms and search heuristics.