States of open quantum systems usually decay continuously under environmental interactions. Quantum Markov semigroups model such processes in dissipative environments. It is known that a finite-dimensional quantum Markov semigroup with detailed balance induces exponential decay toward a subspace of invariant or fully decayed states. In contrast, we analyze continuous processes that combine coherent and stochastic processes, precluding detailed balance. First, we find counterexamples to analogous decay bounds for these processes and prove conditions under which they fail. Second, we prove that the relationship between the strength of local noise applied to part of a larger system and overall decay of the whole is non-monotonic. Noise can suppress interactions that would spread it. Faster decay of a subsystem may thereby slow overall decay. We observe this interplay numerically and its discrete analog experimentally on IBM Q systems. Our main results explain and generalize the phenomenon theoretically. Finally, we observe that in spite of its absence at early times, exponential decay re-appears for unital, finite-dimensional semigroups at finite time.

翻译：开放量子系统的状态通常会在环境相互作用下连续衰减。量子马尔可夫半群在耗散环境中对此类过程进行建模。已知具有细致平衡条件的有限维量子马尔可夫半群会诱导向不变子空间或完全衰减态的指数衰减。相比之下，我们分析了结合相干与随机过程的连续过程，该类过程排除了细致平衡条件。首先，我们发现了这些过程类似衰减界的反例，并证明了其失效的条件。其次，我们证明了施加于大系统局部的噪声强度与整体衰减之间的关系是非单调的。噪声可能抑制其传播的相互作用，因此子系统的更快衰减可能反而减缓整体衰减。我们在数值上观察到了这种相互作用，并在IBM Q系统上通过实验验证了其离散类比。我们的主要结果从理论上解释并推广了这一现象。最后，我们观察到尽管在初始时刻不存在，对于单位性有限维半群，指数衰减会在有限时间内重新出现。