Above the upper critical dimension, the breakdown of hyperscaling is associated with dangerous irrelevant variables in the renormalization group formalism at least for systems with periodic boundary conditions. While these have been extensively studied, there have been only a few analyses of finite-size scaling with free boundary conditions. The conventional expectation there is that, in contrast to periodic geometries, finite-size scaling is Gaussian, governed by a correlation length commensurate with the lattice extent. Here, detailed numerical studies of the five-dimensional Ising model indicate that this expectation is unsupported, both at the infinite-volume critical point and at the pseudocritical point where the finite-size susceptibility peaks. Instead the evidence indicates that finite-size scaling at the pseudocritical point is similar to that in the periodic case. An analytic explanation is offered which allows hyperscaling to be extended beyond the upper critical dimension.Comment: 23 pages, 8 figure

The behavior of complex networks under failure or attack depends strongly on the specific scenario. Of special interest are scale-free networks, which are usually seen as robust under random failure but appear to be especially vulnerable to targeted attacks. In recent studies of public transport networks of fourteen major cities of the world it was shown that these systems when represented by appropriate graphs may exhibit scale-free behavior [C. von Ferber et al., Physica A 380, 585 (2007), Eur. Phys. J. B 68, 261 (2009)]. Our present analysis, focuses on the effects that defunct or removed nodes have on the properties of public transport networks. Simulating different directed attack strategies, we derive vulnerability criteria that result in minimal strategies with high impact on these systems.PACS. 02.50.-r Probability theory, stochastic processes, and statistics -07.05.Rm Data presentation and visualization: algorithms and implementation -89.75.Hc Networks and genealogical trees