Computer system for federated cooperative control of self-learning software agents for secure banking processes in trust-minimized security architectures

Abstract

A computer system for the federated cooperative control of self-learning software agents for secure banking computer processes in trust-minimized security architectures, comprising multiple self-learning software agents arranged in separate banking computer domains, a federated agent coordination layer, a banking process state object creation unit, a trust-minimized access enforcement layer, an agent action arbitration unit, and an output interface, characterized in that the banking process state object creation unit generates dynamic banking process state objects for banking process events, wherein each banking process state object has at least one identity verification level, a device trust value, a session integrity value, a transaction deviation value, an access context value, a policy binding state, an agent trust value, and a process urgency value.includes a risk propagation status and an action authorization state.