Beale Lecture 2023

Operational Research through Seven Decades

Robert Dyson

2023's Beale Lecture, which was held on 16 February, covered Prof Robert Dyson's seven decade career in OR. Beginning in the 60s with work in the glass industry coinciding with a PhD at Lancaster and a first encounter with the professional membership debate. The 70s included a brief encounter with stochastic programming and soft OR which recurred throughout the decades and the start of research into strategy support. The latter continued in the 80s and beyond and work on data envelopment analysis began.  In the 90s, a strong DEA team was built at Warwick, and Robert led the development of the Society's fellowship scheme.  In the 2000s a group met regularly at Warwick leading to a book on strategy support, and EJOR editing began.  In the 10s editing continued and work began on the contribution of the founders of OR to soft OR and practice.  Editing wound down in the 2020s and the founders paper was published in the Journal Operations Research.

Bio

Initially a research mathematician and senior systems technologist at Pilkington Plc (1964-70). Joined Warwick Business School in 1970 as a lecturer. Chair (Head) of WBS 1978-81 and Dean 1998-2000. Visiting Fellow at Technische Hogeschool Twente, Enschede (1977) and Visiting Professor at the University of Texas, Austin (1982). Pro-Vice Chancellor of the University 1989-95 and 1999-2005. Chair of the Committee of Professors of Operational Research (1995-7) and President of the UK Operational Research Society (1998 and 99). Elected a Companion of Operational Research (2007). Board Member of the Coventry Partnership (1998-2005). Governor of Kenilworth School 1989-98, Chair of Governors 94-97. An Editor of the European Journal of Operational Research 2006-2020.  Has published extensively in the fields of cutting and packing, stochastic programming, financial modelling, strategy support, data envelopment analysis and performance management and is a fellow of the OR Society.

Watch Beale Lecture 2023 here

Jake Clarkson

Jake Clarkson is a postdoctoral researcher within NEO team at the Centre Inria d'Université Côte d'Azur. Before joining Inria, he completed his PhD on optimal search for a hidden target in 2020 at the STOR-i Centre for Doctoral Training at Lancaster University, UK. The project was in collaboration with the Naval Postgraduate School in Monterey, California. He then remained at STOR-i for a 1-year postdoc on the management of perishable-product inventory systems. His research interests involve applying probability and mathematical models to real-world problems, currently queuing models.

Doctoral Award Winner Presentation

Optimal search in discrete locations: extensions and new findings

A hidden object needs to be found in many real-life situations which involve large search costs and significant consequences with failure. Therefore, efficient search methods are paramount. In 1962, Blackwell found that a remarkably simple policy minimises the expected search time for an object hidden in one of n discrete locations. This talk will discuss two major extensions to Blackwell’s problem.

The first, motivated by advanced search technology, allows locations to be searched either slowly or quickly, with the faster search using less time but increasing the probability of overlooking the hidden object. Whilst Blackwell’s result may determine which location to search, it falls short of choosing which speed to use. Our main results find conditions on a location to always prefer the same search speed. Outside of these conditions, we propose a heuristic method to choose a speed that demonstrates near-optimal performance in an extensive numerical study.

The second extension replaces the inanimate object with an intelligent hider who aims to maximise the search duration. The resulting two-person zero-sum search game opens up new applications, such as an intruder or computer hacker. Due to the infinite number of pure search strategies, the game is difficult to solve, with the current literature limited to two locations or locations searched in unit time. We extend much of this existing work to the fully-general search game, in particular the existence of an optimal search strategy.