OR and the challenge to improve the NHS: modelling for insight and improvement in in-patient flows

This paper considers efforts to improve in-patient flows, a particularly urgent issue in the National Health Service (NHS). The context is described and related to reasons why OR has been making relatively little contribution. The paper argues that large complex models may often be unnecessary and even get in the way of providing clear insight and guidance for problem owners. The importance of understanding the generic working of systems to lead to improvement, and the limitations of simply describing them, is stressed. It is demonstrated that some very simple models can be of significant practical value in understanding and managing complex systems, changing mindsets and driving collection and use of operationally valuable data. Recommendations for more effective engagement with the NHS are offered.     

What is a design? How should we classify them?

Design theory crosses the boundary between mathematics and statistics, and includes a wide range of disparate types of design. In this paper we present a classification scheme which aims to include as many important types as possible, based on a balance among concept, representation and use.     

Synthesis of perfluorinated functionalized, branched ethers

We wish to report synthesis of perfluorinated functionalized, branched ethers from their hydrocarbon analogues by direct fluorination. Yields up to 90%, with high purities, have been obtained at ambient temperature and pressure. This technique will likely develop into a new general method for producing perfluorinated, hyperbranched and dendritic polymers.     

Embedding partial Steiner triple systems so that their automorphisms extend

It is shown that there is a function g on the natural numbers such that a partial Steiner triple system U on u points can be embedded in a Steiner triple system V on ν points, in such a way that all automorphisms of U can be extended to V, for every admissible ν satisfying ν > g(u). We find exponential upper and lower bounds for g. © 2005 Wiley Periodicals, Inc. J Combin Designs.     

Results of radius scaling experiments and analysis of neon K-shellradiation data from an inductively driven Z-pinch

The K-shell radiated energy (yield) from neon Z-pinch implosions with annular, gas-puff nozzle radii of 1, 1.75, and 2.5 cm was measured for implosion times from 50 to 300 ns while systematically keeping the implosion kinetic energy nearly constant. The implosions were driven by the Hawk inductive-storage generator at the 0.65-MA level. Initial neutral-neon density distributions from the nozzles were determined with laser interferometry. Measured yields are compared with predictions from zero-dimensional (0-D) scaling models of ideal. One-dimensional (1-D) pinch behavior to both benchmark the scaling models, and to determine their utility for predicting K-shell yields for argon implosions of 200 to >300 ns driven by corresponding currents of 4 to 9 MA, such as envisioned for the DECADE QUAD. For all three nozzles, the 0-D models correctly predict the Z-pinch mass for maximum yield. For the 1and 1.75-cm radius nozzles, the scaling models accurately match the measured yields if the ratio of initial to final radius (compression ratio) is assumed to be 8:1. For the 2.5-cm radius nozzle, the measured yields are only one-third of the predictions. Analysis of K-shell spectral measurements suggest that as much as 70% (50%) of the imploded mass is radiating in the K-shell for the 1-cm (1.75-cm) radius nozzle. That fraction is only 10% for the 2.5-cm radius nozzle. The 0-D scaling models are useful for predicting 1-D-like K-shell radiation yields (better than a factor-of-two accuracy) when a nominal (≈10:1) compression ratio is assumed. However, the compression ratio assumed in the models is only an “effective” quantity, so that further interpretations based on the 0-D analysis require additional justification. The lower-than-predicted yield for the 2.5-cm radius nozzle is associated with larger radius and not with longer implosion time, and is probably a result of two-dimensional effects     

On the structure of a random sum-free set

Summary There is a natural probability measure on the set of all sum-free sets of natural numbers. If we represent such a set by its characteristic functions, then the zero-one random variabless(i) are far from independent, and we cannot expect a law of large numbers to hold for them. In this paper I conjecture a decomposition of into countably many more tractible pieces (up to a null set). I prove that each piece has positive measure, and show that, within each piece, a random set almost surely has a density which is a fixed rational number depending only on the piece. For example, the first such piece is made up of sets consisting entirely of odd numbers; it has probability 0.218 ..., and its members almost surely have density 1/4.