A Neural Network Approach to the Sequencing of Construction Tasks

The paper describes the development of a neural network based method for solving complex construction operational problems. A brief introduction to neural networks is provided. Particular reference is made to the mechanics and properties of the technique and its potential as a management problem solving tool. Following this, the optimal sequencing of construction tasks (with the objective of minimizing production time) is selected as an exemplary problem for the application of neural networks. A method of tackling this class of problems, using networks developed through a process of simulated evolution, is proposed. The effectiveness of this approach is then evaluated in terms of both the rate at which networks can be evolved and the efficiency of the solutions they produce. The paper concludes with an indication of areas of current research.