Abstract

Aim: To reveal outcomes obtained with aid the firefighting and rescue response model and area risk analysis for one district.

Introduction: The model, describing responsiveness of the firefighting and rescue system and area analysis of hazards, were described in articles: Spatial analysis of hazards based on historical data and Classification of incidents based on historical data (BiTP Vol. 39 Issue 3, 2015) and, Rescue and firefighting response model (in the current issue of the quarterly - editorial note). This article contains calculations and forecasts derived from the use of aforementioned model, to evaluate the performance of the firefighting and rescue system across one district. Because of content volume constraints, only basic results are included. Nevertheless, these allow for an evaluation of the usefulness of proposed algorithms. The article contains results obtained by the application of standard calculation methods. However, the model was primarily intended for computerised systems, which supported firefighting and rescue planning activity, based on historical data. The authors’ view is that the construction of such systems and consequently their application, would increase the effectiveness of the system and be recognised as an effectiveness increase with limited financial outlay.

Methodology: Analysis, inference and statistical modelling.

Conclusions: Results from the analysis clearly indicate a high level of plausibility in the application of the proposed model to incidents, which have some historical stability, such as: fires, vehicle collisions, road traffic accidents and other local hazards. However, in the case of devastating and very rare incidents; removing the effects of natural catastrophes and large fires, and local threats, the model is difficult to apply because of modest availability of historical data. A significant limitation approximation of empirical distributions accepted by a priori of theoretical distributions. However, considering the capabilities of current computer systems, one can substitute this by neural networks, which, based on historical data, can learn to simulate empirical distributions of individual variables much more accurately. At the same time, obtained results will be more reliable and less burdened by extremes. Results presented in this article clearly demonstrate the usefulness of this model for planning associated with the firefighting and rescue system, even if a desktop application, with principles of the model, is not developed.

Keywords: statistical modelling, statistical analysis, examination of data

Type of article: case study