Least-cost Disjoint Paths with Dependent Cost Structure in Wavelength Continuous Optical WDM Networks

One of the important issues in establishing a fault tolerant connection in a wavelength division multiplexing optical network is computing a pair of disjoint working and protection paths and a free wavelength along the paths. While most of the earlier research focused only on computing disjoint paths, in this work we consider computing both disjoint paths and a free wavelength along the paths. The concept of dependent cost structure (DCS) of protection paths to enhance their resource sharing ability was proposed in our earlier work. In this work we extend the concept of DCS of protection paths to wavelength continuous networks. We formalize the problem of computing disjoint paths with DCS in wavelength continuous networks and prove that it is NP-complete. We present an iterative heuristic that uses a layered graph model to compute disjoint paths with DCS and identify a free wavelength.

Continuous/Cluster-Pinned Recording Media

We propose and theoretically investigate a new class of nanostructured magnetic recording films, cluster-pinned recording media. The films consist of magnetic clusters exchange coupled to a continuous hard layer with perpendicular anisotropy and low coercivity. Our calculations yield the coercivity and the cross-track correlation length as a function of film thickness and pinning density and strength. The mechanism is very similar to the Gaunt–Friedel pinning in bulk magnets, which differs from ordinary strong pinning by the selfconsistent dependence of wall curvature and coercivity on defect concentration. The main difference is the exponent for the coercivity as a function of the pinning strength, which is equal to 2 in the bulk but equal to 3/2 in thin films. The pinning strength is estimated for various regimes, and it is shown that the diminished domain-wall curvature reduces jitter.

Planning “Discrete” Movements Using a Continuous System: Insights from a Dynamic Field Theory of Movement Preparation

The timed-initiation paradigm developed by Ghez and colleagues (1997) has revealed two modes of motor planning: continuous and discrete. Continuous responding occurs when targets are separated by less than 60° of spatial angle, and discrete responding occurs when targets are separated by greater than 60°. Although these two modes are thought to reflect the operation of separable strategic planning systems, a new theory of movement preparation, the Dynamic Field Theory, suggests that two modes emerge flexibly from the same system. Experiment 1 replicated continuous and discrete performance using a task modified to allow for a critical test of the single system view. In Experiment 2, participants were allowed to correct their movements following movement initiation (the standard task does not allow corrections). Results showed continuous planning performance at large and small target separations. These results are consistent with the proposal that the two modes reflect the time-dependent “preshaping” of a single planning system.