The APL phenomenon: An operational research perspective (Q1264098)

The author shows the power of APL as sophisticated mathematical language for the exposition, description and analysis of operations research algorithms. Descriptions of the arithmetic mean problem and of a job scheduling given both in traditional mathematical notation and pidgin ALGOL serve as examples of the shortcomings of the usual notation. In the next section a short introduction of a subset of the APL language is given to prepare the reader for the APL solutions of the following operational research problems: upper-value, lower-value and saddle points of a two-person zero-sum game, the construction of the set of all non- inferior elements of a finite subset S of \(R^ k\), a dynamic programming solution of the standard knapsack problem, a pivot algorithm for a matrix, and last but not least the simplex method for an LP problem in standard form. It is impressive, that - thanks to the elegance and conciseness of the APL language - the author is able to present executable APL-prototypes for all of the above problems together with the problem description on one and a half pages. Finally, the author concludes that since for APL the division of labour between mathematics and programming is minimal, APL provides an economical and expressive notation for operational research problems. APL is recommended for teaching quantitative OR methods and for prototyping algorithms. Although the article gives a carefully prepared and didactic well-made exposition of the virtues of APL, the demonstration is severely hampered by several typesetting errors in the presented APL-code (e.g. p. 151, second APL solution in the top right column, p. 151, third APL solution in the top right column, p. 151, APL solution of taks 2, p. 152, at least 5 errors in the simplex algorithm, p. 152, last APL expression, p. 153 right column bottom).