An Industrial Application of Prolog

Kent Andersson and Torkel Hjerpe

A full-scale industrial application for tunnelling simulation has been developed in Prolog. A typical tunnelling project with three fronts is illustrated below.

The goal of a tunnelling project is to construct tunnels in one or more directions-building a subway is one example. The activities are drilling, charging, blasting, unloading etc. They may interfere with each other between tunnel fronts and resources used may be shared, for example the drilling equipment. The duration of an activity, such as drilling, can not be completely fixed in advance; it varies stochastically but may be simulated by distribution functions.

Atlas Copco, one of the main suppliers of drilling and mining equipment in the world, has invested in a tunnelling simulation system. The output of the system is the expected finishing time for each tunnel front and the amounts of manpower time, machine time, dynamite, concrete, etc spent. By repeating the simulation, a worst-case, a best-case, a mean value and other statistical measures are obtained.

The input to the system consists of facts such as distribution functions, rock classes, machines, etc., but also rules for how to compute time and amounts of resources spent in different tunnels and rock classes.

The tunnelling simulation system is a PC program developed in Prolog, using finite domain constraints. In addition, MS Word, MS Excel and MS Project are used to state facts and rules, to display statistics (of times and resource amounts spent) and to format project plans in Gantt charts. It is a full-scale application, capable of simulating three kilometres of tunnelling on three fronts in five minutes on an ordinary PC. A full-scale prototype of the system has been developed in three man-months.

The short development time illustrates Prolog's efficiency for developing advanced applications. Prolog enables the end-user to change and extend the simulation system, by modifying the rules for time and amount computations. The finite domain constraints extension facilitates problem modelling and reduces computation time.

Computing Science Department, Uppsala University

P.O. Box 311, S-751 05 UPPSALA, SWEDEN

Phone: +46 18 471 00 00, Fax: +46 18 52 12 70

E-mail:{Kent.Andersson,Torkel.Hjerpe}@csd.uu.se