Finding safe routes to school using optimisation methods

Encouraging children to walk to school has become an important issue. Government agencies are therefore keen to reduce child pedestrian casualties associated with the school journey. One way forward is to provide each catchment area with a system of designated routes that minimise road accident risk for pupils on foot. However, there is no accepted modelling procedure for identifying safe routes. This paper reviews the obstacles and assesses the potential for ad hoc solution methods, using the road network surrounding a primary school in north London as a test bed. Solutions close to the optimum were obtained by transforming the network to a simpler form and by modifying the objective function, which effectively linearised the problem. This paper assesses the sensitivity of the results to various assumptions and uncertainties in the data, and concludes with a brief comment on the prospects of finding safe routes on a routine basis at local authority level.