Evaluating navigation safety for harbours in Taiwan: An empirical study

摘要

The purpose of this study is to develop a model to evaluate navigation safety for vessels sailing to Taiwanese commercial harbours, based on a dataset of marine casualties. In order to devise the practical model, linear programming and Data Envelopment Analysis (DEA) are applied to develop a DEA model algorithm. Firsly, a simplified DEA model structure is constructed with one input variable, eleven output variables and four decision-making units (DMUs).Then, a dataset of marine casualties are employed to obtain all variables, as well as to appraise the superiorities of all DMUs. Finally, utilising a DEA solver, the highest navigation safety for the DMUs can be evaluated. Furthermore, an empirical survey of navigation safety in Taiwanese commercial harbours is performed to appraise the systematic approach, ie, the DEA model.The results of this study show that: (I) Keelung and Kaohsiung are the most dangerous harbours; (2) Taiwan's safest commercial harbour is Haulien, followed in ranking by Taichung, Kaohsiung and Keelung; (3) Based on the algorithm of the proposed DEA model, the results can reflect reality.