Landscape Terry Jones Dissertation
A new model of fitness landscapes suitable for the consideration of evolutionary and other search algorithms is developed and its consequences are investigated. Answers to the questions “What is a landscape?”, “Are landscapes useful?”, and “What makes a landscape difficult to search?” are provided. The model makes it possible to construct landscapes for algorithms that employ multiple operators, including operators that act on or produce multiple individuals. It also incorporates operator transition probabilities. The consequences of adopting the model include a “one operator, one landscape” view of algorithms that search with multiple operators. An investigation into crossover landscapes and hillclimbing algorithms on them illustrates the dual role played by crossover in genetic algorithms. This leads to the “headless chicken” test for the usefulness of crossover to a given genetic algorithm and to serious questions about the usefulness of maintaing a population. A “reverse hillclimbing” algorithm is presented that allows the determination of details of the basin of attraction points on a landscape. These details can be used to directly compare members of a class of hillclimbing algorithms and to accurately predict how long a particular hillclimber will take to discover a given point. A connection between evolutionary algorithms and the heuristic search algorithms of Artificial Intelligence and Operations Research is established. One aspect of this correspondence is investigated in detail: the relationship between fitness functions and heuristic functions. By considering how closely fitness functions approximate the ideal for heuristic functions, a measure of search difficulty is obtained. This measure, fitness distance correlation, is a remarkably reliable indicator of problem difficulty for a genetic algorithm on many problems taken from the genetic algorithms literature, even though the measure incorporates no knowledge of the operation of a genetic algorithm. This leads to one answer to the question “What makes a problem hard (or easy) for a genetic algorithm?” The answer is perfectly in keeping with what has been well known in Artificial Intelligence for over thirty years.
Terry Jones is a Founder and Chief Technology Officer of Fluidinfo, a New York based start-up aiming to bring about a fundamental change in how people work with information by building FluidDB, a new kind of online "cloud" database. Previously, from March 2006 to April 2012 Terry was Founder and Chief Executive Officer of Fluidinfo.
Prior to Fluidinfo, from August 2004 to August 2007 Tery was a research associate (i.e., postdoc) in the Department of Zoology at the University of Cambridge (UK).
From 2004 to 2006 Tery was a part-time professor in the Department of Technology (more or less Computer Science) at the Universitat Pompeu Fabra in Barcelona. Terry taught the first-year Ph.D. class on Theory and Algorithms.
Prior to Universitat Pompeu Fabra, Terry was the Chief Technical Officer, acting Chief Operating Officer, and a board member of Eatoni Ergonomics Inc. in New York during 2000-2004.
From 1998 to 1999, Terry was a postdoc at the University of California, San Diego in the Department of Cognitive Science, supervised by Jim Hollan in the Distributed Cognition and Human Computer Interaction Laboratory.
From 1996 to 1998, Terry owned and ran Teclata in Barcelona. Teclata was a small web design and ISP company that Terry co-founded with Ana Mosterín in March 1996. Terry did many things: running the company in general, financing, marketing, pricing, strategy, sub-leasing bandwidth, supporting co-located servers, and hiring. Terry looked after or organized all technical needs: running the web server; generating HTML; writing CGI programs; network security, kernels, PPP, ISDN, FTP, sendmail, pop, named, intranet, a modem pool, etc. In 1998 they sold Teclata to Filnet so Terry could move to UCSD to pursue more research work.
Terry was the Chief Knowledge Engineer for Chiliad, an internet publishing company they launched at the Frankfurt Book Fair in October 1997. Chiliad was based in San Francisco, London and Aix en Provence, with development mainly done in Teclata's offices in Barcelona. Chiliad is now based in Massachusetts. Terry was responsible for the design and development of Chiliad's original product, a networked modular content and authoring system. Terry also evaluated, integrated and customized third-party search engine software, and maintained the company web site.
Terry was a visiting professor in the Dipartimento di Economia Politica at the Università di Modena, Italy. Terry gave two special short courses (8 ninety minute lectures) on Evolutionary Computation in April 1996 and May 1997.
Terry worked as an occasional internet proxy and security consultant for the Junkbusters Corporation in New Jersey. Terry provided low-level advice on HTTP proxy deployment and participated in the design and coding of the Internet Junkbuster, an HTTP proxy for web browsing with a variety of privacy options, including cookie removal, advertisement blocking, customized browser identifiers, and referrer settings.
Terry earned a Ph.D. in Computer Science from the University of New Mexico (1994-1998) with a dissertation on evolutionary computation, fitness landscapes and search, an M.Math from the University of Waterloo (1986-1989) and a B.Sc. (hons) in Computer Science from Sydney University (1982-1985).