Air transportation planning using neural networks as an example of the transportation squadron in the Japan Air Self-Defense Force

The task of a squadron at the Japan Air Self-Defense Force (JASDF) is to transport personnel and materials, and the efficient operation of the squadron is a crucial issue in an emergency. However, transportation is planned manually at present, and it is difficult to quickly formulate an efficient plan. In this context, air transportation planning is considered a transport problem, and an air transportation planning method using a mutually coupled neural network composed of binary neurons, while considering the hardware implementation, is proposed. In the proposed neural network, air transportation planning is partitioned into three subproblems, which are successively solved by three neuron blocks. By this provision, the scale of the problem is reduced and the solution is derived efficiently. Several practical problems with various numbers of aircraft and transportation demands are considered, and simulations are performed using the neural network with various parameters. An examination of the results reveals that the proposed neural network is able to determine whether the air transportation planning can be performed with high speed and high accuracy. In other words, the proposed method is an efficient solution method with many advantages. © 2004 Wiley Periodicals, Inc.