The simulation of gasoline queuing

Abstract

Lines at gasoline service stations have become a fre quent phenomenon in many countries. When lines occur, drivers demand immediate action but government and oil company officials are uncertain about what steps to take. The Battelle event-structured simulation model GASQUE is designed to show which policies are most effective in reducing the length of lines and the amount of time spent in lines.

GASQUE simulates the interaction of 1 500 automobiles and three stations. Many parameters of the model are selected from distributions over time of day, day of week, or the 1 500 automobiles. Such parameters in clude purchase decision levels, consumption rates, driver availabilities, tank sizes, and station hours. The model also permits the simulation of end-user allocation plans such as assigning purchase days by automobile license number (e.g., the so-called odd- even plans). Output from the model includes statis tics such as average line length and average number of hours spent in line per week.

The application of GASQUE requires the use of compara tive analysis, that is, the examination of changes from standard or base-case simulations. In one series of simulations, studying the joint effects of tank-topping (buying a small amount of gasoline sim ply to keep the tank full) and reductions in station hours shows that either may cause lines to form. Imposing a maximum purchase limitation causes longer lines, especially when combined with reduced hours.

A second series of simulations studied the effects of end-user allocation plans. In general, lines are made shorter by requiring minimum purchases and re quiring stations to remain open on weekends. Odd- even plans result in longer lines. A comparison of these findings with reports from various sections of the United States suggests that the simulation results agree with the 1979 experience with gasoline queues.