Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


The purpose of this work is twofold: (1) to create a compre­hensive and complete data processing system specifically designed to permit efficient access to specific information contained in traffic accident information files, and (2) to demonstrate an appli: :ion of the system using a model developed for the optimal allocation of per­sonnel to reduce accidents by means of selective law enforcement. The Generalized Data Extraction System consists of two totally integrated programs: (1) the Accident Data File Creation Program, and (2) the Generalized Data Extraction Program. The Accident Data File Creation Program extracts information from a master accident data file, edits the information, and creates a reformatted permanent Accident Data File. The objective is to perform all those operations in file creation which would be time consuming and inefficient if performed during the search routines conducted during data extraction. The Generalized Data Extraction Program is specifically designed to interface directly with the Accident Data File and to be utilized by persons not trained in computer programming. In essence, this program permits the user to 'ask the computer a question' and obtain the answer in the form of three different formats the program is capable of generating. The three output formats the program is capable of generating are: (1) an accident swmnary report, (2) a high accident location report, and (3) a histogram depicting the accident experience for given stretches of roadway. To demonstrate an application of the Generalized Data Extrac­ tion System, a model is developed which provides for the optimal assignment of law enforcement personnel to maximize the reduction in the number of violations which have resulted in accidents subject to the constraints imposed by the scarcity of the resource and other relevant factors. Using a combination of concepts developed in this work and existing computer hardware/software systems, a total real-time program with the specific objective of optimally reducing accidents by timely and appropriate assignment of personnel has been discussed. The feasibility and efficiency of the Generalized Data Extrac­tion System and the concepts underlying the system have been successfully demonstrated. While the original system was designed to accommodate vehicle accident data, the underlying concepts extend far beyond this one specific application. Use of these concepts can provide the researcher or the manager with specific information neces­sary to analyze specific problems without resorting tu more costly and time consuming techniques.