Incorporating Socioeconomic Factors into the Facility Location Optimisation Problem in a Mixed Urban-Rural Population with Application to Queensland Fire Stations

The crucial role of emergency services lies in their life-saving interventions, demanding an optimal and effective allocation of resources. Paramount to this efficacy is the strategic physical placement of these resources, ensuring they are ideally situated to meet the fluctuating demands of emergency scenarios. Central to the optimisation process is the concept of the covering location model, a critical analytical tool in planning for the appropriate facility locations.

Initial literature reviews indicate a significant gap: existing models often overlook the impact of socioeconomic variables. Nonetheless, multiple studies have highlighted that socioeconomic factors contribute significantly to the variation in emergency services demand. Changes in the socioeconomic fabric of a population can render location analyses based on historical data inadequate or obsolete. Moreover, there is a conspicuous lack of research on covering location models that specifically address populations consisting of urban and rural subgroups, particularly in regions characterised by sparsely distributed residents.

In light of these findings, this thesis proposes an innovative framework. This comprehensive approach not only integrates crucial socioeconomic variables but also strategically maximises coverage while minimising the distance between areas of uncovered demand and their nearest emergency facilities. The proposed framework is tripartite: it consists of robust backward stepwise regression analysis, advanced covering location modelling, and the application of a random search algorithm.

The initial phase involves a robust backward stepwise regression model, meticulously analysing the correlation between demand for emergency services and various socioeconomic indicators. Following this, a dual-objective optimisation model is employed, leveraging the sophisticated Implicit Modified Coverage Location Problem (MCLP-Implicit) methodology, to pinpoint the most advantageous allocation of fire stations. Subsequently, a random search algorithm, known for its efficiency, is utilised in solving the intricate covering location model.

To validate the framework's applicability, the research conducted an exhaustive case study in South East Queensland. Data critical to the research were obtained from reputable sources, including the Queensland Government and the Australian Bureau of Statistics. The results provide compelling justification for the model's adoption. Notably, the robust backward regression analysis identified sixteen influential variables within the Index of Relative Socioeconomic Advantages and Disadvantages (IRSAD), directly correlating with the incidence of building fires in the region.

The optimisation process, informed by demand data extrapolated from the socioeconomic model, yielded solutions with an expanded coverage area, particularly when assuming a service radius of 7.5 km and 10 km for each facility. This augmented range is indispensable for the dynamic deployment of personnel in response to demand surges, whether due to seasonal trends or unexpected disasters. Significantly, by integrating socioeconomic data into demand forecasting, the model ensures that vulnerable or disadvantaged communities are not neglected.

The insights garnered from this research are invaluable for enhancing long-term urban planning and formulating nuanced fire response protocols, particularly for areas with similar socioeconomic dynamics or disaster risk profiles. Beyond its academic contributions, the model holds substantial commercial value for entities servicing fire and rescue authorities, underscoring its broad-spectrum relevance and potential for real-world application.