Gwilym M. Jenkins

August 12, 1932 – July 10, 1982

Brief Biography

Gwilym Meirion Jenkins was a Welsh statistician and pioneer in time series analysis. Born in Gowrton, Swansea and not speaking English until his was six years old, Jenkins was known for his unusual inventiveness and intuition. He studied mathematics at University College London, receiving his bachelors in 1953 and PhD in 1956. His first two postdoctoral years were spent as a junior fellow at the Royal Aircraft Establishment at Farnborough in Southern England. Tasked with aircraft development, Jenkins focused on undercarriage design. In this capacity, he developed a method for measuring and recording the ups and downs of an aircraft’s wheels as a time series.

The British statistician community in the late 1950s and 1960s was strongly divided between theory and application. Jenkins was an avid proponent against this partition, striving to unite the two forces. Nearly all of his theoretical contributions were tied to real world practice. In 1959, Jenkins made his first visit to the United States sharing his early work on time series analysis as a guest lecturer to Stanford University and Princeton,. It was during this visit to the Statistical Techniques Research Group at Princeton when he first met English-born probabilist George E. P. Box.

This initial meeting between Jenkins and Box led to a sustained and fruitful collaboration between the two statisticians. Between 1960 and 1970, the duo began thinking about automatic optimization and the importance of dynamics. Over time, they started to work more on general discrete control problems, trying to better understand the relation of their time series work to other kinds of statistical control. Jenkins was drawn to the principle that if some method or solution has withstood the test of time, then there must be a good reason for it. In the mid-1960s, he decided that he and Box write a book.  The resulting publication, Time Series Analysis, Forecasting, and Control (1970), was given an honorable mention by the Frederick W. Lanchester Prize committee for best publication in operations research.

Jenkins was made a Reader at Imperial College London in 1964, though shortly after left for Lancaster University’s Department of Systems Engineering. It was during this transitionary period when he was diagnosed with Hodgkin’s lymphoma, the disease which would eventually take his life just weeks before his fiftieth birthday. At Lancaster, he founded and became managing director of the International Systems Corporation of Lancaster. Jenkins also established the Journal of Systems Engineering in 1969.

In 1974, Jenkins left academia to start his own consulting firm, Gwilym Jenkins and Partners. With it, he published two books on time series analysis, Practice Experience with Modeling and Forecasting Time Series (1978) and Case Studies in Time Series Analysis (published posthumously 1983). Among his many honors, Jenkins was an elected member of the Royal Statistical Society, the Institute of Mathematical Statistics, and the Institute of Statisticians. In the mid-1970s, he briefly worked with the Royal Treasury. Never one to shun his creative impulses, Jenkins was an accomplished pianist and jazz enthusiast. 

Other Biographies

Wikipedia Entry for Gwilym Jenkins

Box G. E. (1983) Gwilym Jenkins, experimental design, and the time series. Questiio, 7(4): 515-525. (link)

Mills T. C. (2013) Jenkins: inference in autoregressive models and the development of spectral analysis. in A Very British Affairs: Six Britons and the Development of Time Series Analysis During the 20th Century, 140-160. Palgrave Macmillan: New York. 

Education

University College London, BS 1953

University College London, PhD 1956 (Mathematics Genealogy)

Affiliations

Academic Affiliations
Non-Academic Affiliations
  • Royal Aircraft Establishment

Key Interests in OR/MS

Methodologies

Obituaries

Box G. E. P. (1983) G. M. Jenkins, 1933-1982. Journal of the Royal Statistical Society, 146(2): 205-206.

Awards and Honors

Frederick W. Lanchester Prize 1970

Selected Publications

Jenkins G. M. (1954) Tests of hypothesis in the linear autoregressive model, I. Biometrika, 41(3/4): 405-419.

Jenkins G. M. (1956) Tests of hypothesis in the linear autoregressive model, II. Biometrika, 43(1/2): 186-199.

Box G. E. P. & Jenkins G. M. (1962) Some statistical aspects of optimization and control. Journal of the Royal Statistical SocietySeries B, 24(2): 297-343.

Box G. E. P. & Jenkins G. M. (1963) Further contribution to adaptive quality control: simultaneous estimation of dynamics: no-zero costs. Bulletin of the International Statistical Institute, 34: 943.

Chanmugam J. & Jenkins G. M. (1963) Optimum experimentation in the process industries. Proceedings of the American Institute of Chemical Engineering Symposium, 59: 108-117.

Bacon D. W., Box G. E. P.,  & Jenkins (1967) Models for forecasting seasonal and nonseasonal time series. Harris B., ed. in Advanced Seminar on Spectral Analysis of Time Series, 271. John Wiley & Sons: New York.

Jenkins G. M. & Watts D. G. (1968) Spectral Analysis and its Applications. Holden-Day: San Francisco. 

Box G. E. & Jenkins G. M. (1970) Time Series Analysis, Forecasting, and Control. Holden-Day: San Francisco.

Jenkins G. M. & Youle P. V. (1971) Systems Engineering. Everyman's Library: London. 

Jenkins G. M. (1978) Practice Experience with Modeling and Forecasting Time Series. Gwilym Jenkins and Partners: Lancaster, UK. 

Jenkins G.. M. & McLeod G. (1983) Case Studies in Time Series Analysis. Gwilyn Jenkins and Partners: Lancaster, UK.