Common use of Conclusion and Future Work Clause in Contracts

Conclusion and Future Work. ‌ We revisited Attack-Defence Trees under a unified syntax, extending the usual constructs with a new sequential counter-operator (SCAND). More importantly we introduced XXXXX, an agent-aware formalism to model ADTrees, and trans- formation patters from the latter to the former that are sound, complete, and preserve the compositionality of ADTrees, naturally covering cases with shared subtrees. The impact of different agent coalitions on attack time and cost was evaluated using Uppaal and IMITATOR. Finally, the feasibility of an attack was evaluated through parameter synthesis with IMITATOR, to obtain the attribute values of ADTree nodes that make an attack succeed. Our experiments show that (and how) different agent distributions affect the time of attacks/defence xxxxxx- xxxx, possibly rendering some infeasible. We expect this will open the gate to richer studies of security scenarios, with multiple agents that can collaborate. Our next goals include logics to express properties in EAMAS, and adapting the partial order reduction from [17] as well as the state space reduction for tree topologies of [21] to agent strategies in EAMAS, including extensions to parametric timing information. This will allow for studying the strategic abilities of agents, ultimately in a parametric setting. Finally, we will add support for agents assignment to our tool adt2amas that transforms ADTrees into EAMAS. References

Conclusion and Future Work. ‌ We revisited Attack-Defence Trees under a unified unified syntax, extending the usual constructs with a new sequential counter-operator (SCAND). More importantly we introduced XXXXX, an agent-aware formalism to model ADTrees, and trans- formation patters from the latter to the former that are sound, complete, and preserve the compositionality of ADTrees, naturally covering cases with shared subtrees. The impact of different different agent coalitions on attack time and cost was evaluated using Uppaal and IMITATOR. Finally, the feasibility of an attack was evaluated through parameter synthesis with IMITATOR, to obtain the attribute values of ADTree nodes that make an attack succeed. Our experiments show that (and how) different different agent distributions affect affect the time of attacks/defence xxxxxx- xxxx, possibly rendering some infeasible. We expect this will open the gate to richer studies of security scenarios, with multiple agents that can collaborate. Our next goals include logics to express properties in EAMAS, and adapting the partial order reduction from [17] as well as the state space reduction for tree topologies of [21] to agent strategies in EAMAS, including extensions to parametric timing information. This will allow for studying the strategic abilities of agents, ultimately in a parametric setting. Finally, we will add support for agents assignment to our tool adt2amas that transforms ADTrees into EAMAS. References