Applications of Evolutionary Computation

In directed networks, flow dynamics may lead to cascade failures due to node and link removal. The systemic risk in financial systems follows similar mechanism, where banks are connected by interbank linkages with money transfers. A mathematical model of the banking network is used to investigate the relationships between the cascade dynamics and key parameters determining the banking network structure, including the connectivity, the bank’s capitalization, and the size of interbank exposure, based on analytical calculations and numerical simulations. To optimize the network topology for the minimization of systemic risk, genetic algorithm is applied to evolve the network. It is observed that the systemic risk of financial system could be decreased by increasing the degree variance of the associated network. This could be useful for financial risk management, with possible applications to other physical systems such as ecological web, where the network stability is also an important issue.