Study of fleet assignment problem using a hybrid technique based on Monte Carlo simulation and genetic algorithm

Abstract

Fleet assignment problem (FAP) is the assignment of an aircraft model to each scheduled flight based on key operational variables such as cost, revenue, passenger travel demand and aircraft specifications. FAP is an important aspect of aircraft planning within an airline. While many developed economy have automated this planning task, developing economy such as Nigeria mainly depend on manpower to carry out this task. The aim of this paper is to solve a FAP using a hybrid technique based on the combination of Monte-Carlo (MC) simulation and Genetic Algorithm (GA). The objective function is total cost and variation in aircraft models and passenger traffic associated with different scheduled flight were considered. MC simulation which was carried out based on the numerical approximation of normal distribution cumulative distribution function (cdf) was used to estimate the expected passenger spill rate, while genetic algorithm was used for the optimization. The result was found to be satisfactory, as optimal fleet plan was achieved in approximately fifteen seconds of program run time, as against not less than an hour usually spend using human effort to solve FAP. Also the optimized plan resulted to a thirty percent saving in comparison to the actual plan implemented by the airline. It is therefore recommended that MC-GA optimization technique should be considered as an alternative technique applicable for FAP optimization.

Keywords: Fleet assignment, genetic algorithm Monte-Carlo simulation, optimization