Abstract | ||
---|---|---|
The current hierarchical, fossil-fuel based energy system is shifting towards a sustainable system based on distributed renewable generation. Simultaneously, energy end consumers become increasingly important as active prosumers. Local electricity markets (LEMs), on which prosumers and consumers can trade electricity locally, enable sustainable, distributed local electricity balances with an active involvement of the end customers. However, trading needs to be automated, and specified to the household's specific preferences in terms of price and electricity source. We show how intelligent agent strategies can fulfill both objectives. To this end, we conduct a multi-agent simulation of a LEM between 100 households and a community storage in a merit order LEM. LEM agents maximize their individual utility via automated Erev-Roth reinforcement learning. The learning strategies take into account the households' individual electricity preferences. To this end, agent preferences are grouped into truly greens, price-sensitive greens, and non adopters. The evaluation of the strategies is based on the agents' revenues, costs and electricity source mix. It shows that reinforcement learning can represent household preferences on LEMs. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1145/3208903.3214348 | E-ENERGY'18: PROCEEDINGS OF THE 9TH ACM INTERNATIONAL CONFERENCE ON FUTURE ENERGY SYSTEMS |
Keywords | Field | DocType |
Local electricity market,intelligent agents,reinforcement learning,household agents | Revenue,Intelligent agent,Electricity,Renewable generation,Energy system,Merit order,Cluster analysis,Environmental economics,Business,Reinforcement learning | Conference |
Citations | PageRank | References |
0 | 0.34 | 4 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Esther Mengelkamp | 1 | 35 | 3.90 |
Christof Weinhardt | 2 | 985 | 141.98 |