Networks must be resilient to challenges such as malicious attacks or network overload and adapt their operation in an autonomous manner. Network simulations enable the testing of complex network scenarios (which would be difficult to emulate using actual hardware) in an inexpensive manner. However, it is difficult to evaluate resilience strategies that involve the interplay between a number of detection and remediation mechanisms that must be activated on demand according to events observed in the network (as opposed to hardcoded protocols). In this paper we propose the notion of a policy-based resilience simulator based on the integration of a network simulator and a policy management framework. This permits the evaluation of resilience strategies consisting of mechanisms whose behaviour can be adapted during run-time - e.g. setting flags, dropping connections, triggering or stopping monitoring sessions, etc. We employ policies to specify the required adaptations, which are de-coupled from the hard-wired implementations of the simulated components, according to conditions observed during run-time in the simulation. We can thus observe how real policies affect the operation and the behaviour of simulated components, and then evaluate the effectiveness of resilience strategies before they are deployed in the network infrastructure.