Updated: Dec 27, 2021
by Hans Ittmann
Gerhard J Rudolph, who died in Pretoria on 15 January 2021, played a huge role in the early years of the Operations Research Society of South Africa (ORSSA) and was instrumental in introducing Operations Research (OR) at UNISA in 1968. This obituary is based partly on personal recollections, interactions with Gerhard’s wife and information gathered from a variety of ORSSA sources.
The soft-spoken Gerhard was born in Rustenburg on 8th January 1937. He matriculated from Ermelo High School in 1953 with distinctions in Latin, Mathematics and Accounting. He initially enrolled for a BA degree with Latin, Greek, Economics and Mathematics at the University of Pretoria, but he changed to a BSc and graduated in 1957 with a distinction in Mathematics. In 1956, he was introduced to Mathematical Statistics, thus participating in the first such course presented at the university. Mixing studies and work, first at Unisa and then at the CSIR, he obtained an MSc and a PhD, the latter completed in 1963. The title of his PhD thesis was “Oor eksakte en asimptotiese benaderde statistiese toetse by geklassifiseerde gegewens”.
It was during this time that he had the privilege to work on what was possibly the first scientific computer in South Africa, that the CSIR acquired. This was the ZEBRA computer (Zeer Eenvoudig Binêre RekenAutomaat) that provided many hours of fascination and frustration, not only for Gerhard, but also a few young scientists, many of whom later became involved in OR. The unit at the CSIR where Gerhard worked became the National Research Institute for Mathematical Sciences (NRIMS) and it was here that the group was exposed to OR. In 1962/63 Gerhard Geldenhuys was requested by the CSIR to conduct a survey of possible applications of OR and to present a series of lectures on the topic to NRIMS staff. Soon afterwards Gerhard and two colleagues, joined by their wives, went to the UK to familiarise themselves further with the exciting new field of OR.
computer in South Africa, that the CSIR acquired. This was the ZEBRA computer (Zeer Eenvoudig Binêre RekenAutomaat) that provided many hours of fascination and frustration, not only for Gerhard, but also a few young scientists, many of whom later became involved in OR. The unit at the CSIR where Gerhard worked became the National Research Institute for Mathematical Sciences (NRIMS) and it was here that the group was exposed to OR. In 1962/63 Gerhard Geldenhuys was requested by the CSIR to conduct a survey of possible applications of OR and to present a series of lectures on the topic to NRIMS staff. Soon afterwards Gerhard and two colleagues, joined by their wives, went to the UK to familiarise themselves further with the exciting new field of OR.
The group spent a month at the University of Manchester attending some OR courses and then joined the British Army Operational Research Establishment (AORE) where they worked jointly with OR scientists of the unit. Afterwards Gerhard and his wife went to the USA where Gerhard spent a semester at Stanford University in California. He did a course in dynamic programming under the well-known Prof. Ken Arrow. They returned to South Africa in 1965 after a five-week vacation in Europe.
In 1967 Gerhard left the CSIR, accepting a senior lectureship at UNISA in the Department of Statistics. His focus was to get an MSc programme in OR in place. The course was introduced in 1968 and there were 60 students on that first course. This again led to Gerhard accepting a Chair in Statistics at Rhodes University in 1969 with the objective of building the department from scratch. It was during 1969 that Gerhard was recruited by the then deputy Postmaster-General, Mr Louis Rive. Rive was the previous Secretary of Defence and he was fully aware of the possibilities and potential benefits of OR in the postal environment. The Rudolphs relocated to Pretoria in 1970 for Gerhard to take up his new position as head of an operations research group he was to establish in the Department of Posts and Telegraphs. Gerhard remained with this group at the Postal Services, which later formed part of Telkom, till his retirement in 1993.
Having been exposed to the new field of OR Gerhard not only launched himself wholeheartedly into acquiring further skills in OR and then into teaching OR, but he also assisted and helped guiding ORSSA through its early days. He was one of the founding members of the Society and at the two-day founding conference in November 1969 presented a paper titled “A model for forecasting – estimating footwear production in South Africa”. Being in the Eastern Cape where there were a substantial number of shoe factories, Gerhard clearly started contributing to this industry almost immediately.
During the founding meeting Gerhard was elected as one of the two additional members of the first ORSSA Executive. This started a long period of service to the Society. At the next conference Gerhard was elected as treasurer of the second executive committee. The following year he became the vice-president of the Society and in 1973 served as the fourth ORSSA president. The annual conference that year was the first conference I attended, and I remember vividly how formal the occasion was when he delivered his presidential address. Gerhard’s wife was also in attendance. His address emphasised that the benefit of “O.R. lies in applications in industry, business, government and defence. To achieve this benefit, O.R. must grow in breadth as well as in depth”. Even at that early stage of my involvement in OR it resonated well with me. Gerhard also served as chairman of the Pretoria chapter in subsequent years.
In 1971 Gerhard was the first recipient of the recently instated Tom Rozwadowski medal. His paper “Optimal mail sorting policy” published in the ORSSA newsletter in June 1971 was a worthy winner of this award. I think it is appropriate to elaborate on this work. Mail sorting constituted one of the major 4 000 man-years of labour per annum. Any small percentage improvement in sorting procedures would therefore have resulted in a considerable absolute reduction of labour costs and an alleviation of postal delays. The problem addressed was the following: mail, in the form of letters or parcels, had to be sorted for dispatch or delivery to n possible destinations. Equipment, either in the form of a sorting rack or a sorting machine, comprising r sorting bins was used where r < n. What needed to be determined was in what manner and sequence the sorting bins should be allocated to single destinations and to groups of destinations to minimise sorting time and labour requirements. Previous mail sorting methods and policies were studied, and it was realised that for an optimal mail sorting policy one could not determine sorting rules in advance because these rules themselves depended on the mail distribution. Mail distribution should therefore rather be used for determining an optimal mail sorting policy. A dynamic programming approach was developed to solve this problem which led to considerable savings both in terms of delays and labour costs. Having attended a course in dynamic programming at Stanford, I can imagine that Gerhard was keen to find a problem where he could apply this technique. As it turned out this was the ideal problem!
As Dave Masterson, Past-President of ORSSA, stated in a piece on Gerhard some years ago, Gerhard revived his own interest in economics. This was typical of his approach to the application of OR – whilst one can manage and determine one’s own objectives, it is tremendously important to understand the underlying economics of the situation and to understand the management of the non-academic environment of the problem domain.
At the annual conference in 1999 at Dikhololo, celebrating the 30th anniversary of ORSSA, the society bestowed honorary membership on Gerhard, recognising him for his valuable contribution to OR and the Society of which he was a founding member.
Gerhard and his wife shared the same interest in Statistics and OR. Heilet worked in the OR and Statistics group at the CSIR for many years. Their eldest son, Mynhard, obtained a PhD in Mathematics while their second son, Gerhard, is a successful attorney in Johannesburg. The two boys both excelled in fencing as youngsters. Mynhard is married to Yvonne Walus, also a PhD in Mathematics. They moved to New Zealand several years ago and have two children. Yvonne was a member of ORSSA and served the Society as chairperson of the Pretoria chapter as well as editor of ORiON. Heilet, Mynhard and Yvonne at some stage all worked in the same group at the CSIR.
After his retirement Gerhard and Heilet did a lot of travelling. This included six visits to their kids and grandchildren in New Zealand. They also visited Eastern Europe and had a delightful boat trip in Russia with two friends. Gerhard, with his knowledge of Greek and Latin, managed to get a basic knowledge and understanding of Russian which was very handy during that vacation. There was also a wonderful tour to China.
John Hearne, ORSSA President on the left, presents Gerhard Rudolph, right, with honorary membership at the annual ORSSA conference in 1999 Gerhard suffered from heart problems and high blood pressure, and over the last couple of months his condition deteriorated. He is survived by Heilet, to whom he was married for 61 years, their two sons and their wives as well as five grandchildren. Our heartfelt thoughts go out to them in this time of sorrow. He will be sorely missed.