---

This thesis builds upon work by Breuer et al. which asked how the leverage cycle model of John Geanakoplos can be implemented in a continuous double auction and under which circumstances holdings and trading prices approach an equilibrium. A main criteria is the trading network the agents use where the authors examine only two topologies and report that the prices come to an equilibrium only in the case of a fully connected network. They leave the question open on how the model behaves with different kinds of networks and which network topology allows an equilibrium to be reached for further research. This thesis builds upon this model and gives a hypothesis for the necessary property a network must satisfy to allow the model to approach the theoretical equilibrium as reported by Breuer et al. A few network topologies are examined in regard of their ability to allow equilibria to be reached or not through agent-based simulation. The results are surprising as they show that the model is incomplete and needs to be extended by an additional market. Although adding this market shows that some problems can be overcome, it becomes clear that the original hypothesis is infeasible.

---

Thaler, Jonathan
Jonathan Thaler, M.Sc.: studies of computer graphics at University of Technology Vienna, complex systems and agent-based simulation at University of Applied Sciences Vorarlberg. Ph.D. student at University of Nottingham researching pure functional programming in Agent-Based Simulation.

---