The issue of automatic price optimization in e-commerce is increasing in importance. This is in particular due to the fact that significant increases in margins can be achieved using intelligent pricing strategies. In addition to the standard algorithms basically designed for optimizing the prices for each product in an online shop, special algorithms are developed that are used for such matters as the profit-oriented sale of combined products or the rapid selling off of perishable goods. The time has now come for the issue of automatic price optimisation to become the focus of the DMC task. In 2012 this again comprises an offline and an online part that are assessed independently of each other.
In the offline part, the first task, the prices and associated sales figures are stated for a specific period and for selected products. With the aid of this data a model is to be developed that describes the dependencies between the data. Based on this model, the sales figures of the subsequent period have to be forecasted in an application phase. The winner of the first task is the team that predicts the sales figures as precisely as possible.
The challenge for the second task is to implement an agent that takes on the pricing role for an online shop. The agents and therefore the shops of the individual participants finally compete against each other in a multi-agent system simulation where they aim to maximize their profits. Therefore, the market demand is simulated by the prudsys AG. The winner is the team that achieves the highest profit.
Task 1 First Place:
Universität Duisburg-Essen, team 1(EUR 2,000.00)
Task 1 Second Place:
Technische Hochschule Mittelhessen, team 1 (EUR 1,000.00)
Task 1 Third Place:
Karlsruher Institut für Technologie, team 2 (EUR 500.00)
Task 2 First place:
Amirkabir University of Technology, team 2
Task 2 Second Place:
Amirkabir University of Technology, team 1
Task 2 Third Place:
Brigham Young University, team 2