No abstract

Location-allocation models simultaneously locate systems of facilities and allocate demand to them. Much of the literature deals with location on networks. Demand is almost always expressed as a weight at nodes: the goal of location is to serve the demand at these nodes, which is allocated to facilities. Based upon the concept of a maximum specified distance beyond which people are not covered by a facility, two familiar models aim at covering demand nodes (Church and Meadows 1979). The set covering location model locates facilities to minimize the number of facilities required to cover all demand. The maximal covering location model (MCLM) locates a specified number of facilities to cover maximum demand. Demand in a network is not always expressed at nodes, however. The retail marketing literature (Snow and Scott 1984; Mercurio 1984; Jones and Mock 1984; Fairbairn 1984) indicates that, for some types of business, traffic flows exert a demand. Examples are convenience stores, gasoline stations (Goodchild and Noronha 1987), banking machines, and other outlets selling impulse items (Ghosh and McLafferty 1987). Other types of facilities, such as billboards, are located for exposure to traffic. Current methods of locating retail facilities to serve traffic simply evaluate locations on the basis of passing flows, determined by traffic counts. This fails to deal with a potentially damaging self-competition problem, termed cannibalization (Ghosh and McLafferty 1987). This term usually applies to locating facilities so close to each other that they cut into each other's market areas. Mathematical programming formulations of the MCLM and other location-allocation models deal conceptually and operationally with multiple facility location, avoiding cannibalization by "coordinat(ing) the location of the entire network and not just evaluat(ing) outlets individually" (Ghosh and McLaf€erty 1987). Flows in networks, as well as market areas, can be cannibalized. Flows passing along links or through nodes are coming from origins and going to destinations; they have passed, and will pass, along other links and through other nodes. Just as demand at nodes may be covered by several facilities located close together, traffic in a network may be served by several facilities located on common paths. Just as multiple covering may lead to other demands going uncovered, multiple traffic Erasmus University, ran C. Amrhein and to N. Bregman at the Ontario Centre for Large Scale Computation for their help. G. Lester, R. Dunphy, and R. Pakan prepared the diagrams. Great thanks to 0. Berman for permissions to use his network. This research was supported by NSERC Canada. Thanks to ohn Current who argued endless hours with me. Ken Rosin the BLP, an d provided welcome comment. The referees made many use% suggestions. Thanks to M. John Hodgson is professor of geography, The University of Alberta.

Neighbourhood spatial accessibility (NSA) refers to the ease with which residents of a given neighbourhood can reach amenities. NSA indicators have been used to inform urban policy issues, such as amenity provision and spatial equity. NSA measures are, however, susceptible to numerous methodological problems. We investigate one methodological issue, aggregation error, as it relates to the measurement of NSA. Aggregation error arises when, for the purpose of distance calculations, a single point is used to represent a neighbourhood, which in turn represents an aggregation of spatially distributed individuals. NSA to three types of recreational amenities (playgrounds, community halls, and leisure centres) in Edmonton, Alberta, Canada is used to assess whether aggregation error affects NSA measures. The authors use exploratory spatial data analysis techniques, including local indicators of spatial association, to examine aggregation-error effects on NSA. By integrating finer resolution data into NSA measures, we demonstrate that aggregation error does affect NSA indicators, but that the effect depends on the type of amenity under investigation. Aggregation error is particularly problematic when measuring NSA to amenities that are abundant and have highly localized service areas, such as playgrounds. We recommend that, when analyzing NSA to these types of amenities, researchers integrate finer resolution data to indicate the spatial distribution of individuals within neighbourhoods better, and hence reduce aggregation error.

Assessing spatial equity with respect to urban public amenity provision involves examining the association between amenity distribution and population need for amenities. Geographic Information Systems in coordination with local spatial autocorrelation, were used to investigate the association between neighbourhood accessibility to playgrounds and demographic and social need for playgrounds in Edmonton, while considering differences in playground quality throughout the city. The primary objectives of this study were to assess whether playground provision, for location and quality, in Edmonton is equitable and, more generally, to investigate the role that amenity quality plays in assessing spatial equity. The results indicate that playgrounds are equitably distributed within Edmonton, with the highest‐social‐need neighbourhoods having the greatest accessibility to playgrounds. However, once differences in playground quality are considered, there is less of an association between high‐social‐need and high‐accessibility areas. The findings suggest that greater attention be paid to differences in playground quality within Edmonton and that spatial equity researchers give greater consideration to amenity quality when evaluating spatial equity within cities.

The nature of discomfort and level of exertion associated with wearing respiratory protection in the health care workplace are not well understood. Although a few studies have assessed these topics in a laboratory setting, little is known about the magnitude of discomfort and the level of exertion experienced by workers while they deliver health care to patients for prolonged periods. The purpose of this study was to determine the magnitude of discomfort and level of exertion experienced by health care workers while wearing respiratory protection for periods up to 8 hr when performing their typical occupational duties. This project was a multiple cross-over field trial of 27 health care workers, aged 24-65, performing their typical, hospital-based occupational duties. Each participant served as his/her own control and wore one of seven respirators or a medical mask for 8 hr (or as long as tolerable) with interposed doffing periods every 2 hr. Self-perceived discomfort and exertion were quantified before each doffing: self-perceived level of discomfort using a visual analog scale, and self-perceived level of exertion using a Borg scale. Overall, and as would be expected, discomfort increased over time with continual respirator use over an 8-hr period. Interestingly, exertion increased only marginally over the same time period. The relatively low level of exertion associated with eight respiratory protective devices, including models commonly used in the U.S. health care workplace, is not likely to substantially influence workers' tolerability or occupational productivity. However, the magnitude of discomfort does appear to increase significantly over time with prolonged wear. These results suggest that respirator-related discomfort, but not exertion, negatively influences respirator tolerance over prolonged periods. Discomfort may also interfere with the occupational duties of workers.