A model of the UK market in electricity combining key factors influencing generator bidding is proposed and a hierarchical multi-objective adaptive agent architecture using case based reasoning and learning classifier systems is described. Experimentation shows that the adaptive agents learn bidding strategies that have been observed in the real world, and that in some market scenarios the agents appear to be learning the benefits of cooperating to receive increased long term rewards. The potential of the adaptive agent model is illustrated by experimentation with an alternative market structure.
|Number of pages||8|
|Publication status||Published - Jul 2000|
|Event||Genetic and Evolutionary Computation Conference (GECCO-2000) - Las Vegas, United States|
Duration: 8 Jul 2000 → 12 Jul 2000
|Conference||Genetic and Evolutionary Computation Conference (GECCO-2000)|
|Period||8/07/00 → 12/07/00|