@inproceedings{eurogp06:TsangJin, title = {Incentive Method to Handle Constraints in Evolutionary}, address = {Budapest, Hungary}, author = {Edward Tsang and Nanlin Jin}, booktitle = {Proceedings of the 9th European Conference on Genetic Programming}, editor = {Pierre Collet and Marco Tomassini and Marc Ebner and Steven Gustafson and Anik\'o Ek\'art}, month = {10 - 12 April}, pages = {133--144}, publisher = {Springer}, series = {Lecture Notes in Computer Science}, url = {http://link.springer.de/link/service/series/0558/papers/3905/39050133.pdf}, volume = {3905}, year = {2006}, biburl = {http://www.bibsonomy.org/bibtex/2a0a07a4f07e977f4c4397bbf444dec09/brazovayeye}, abstract = {This paper introduces Incentive Method to handle both hard and soft constraints in an evolutionary algorithm for solving some multi-constraint optimisation problems. The Incentive Method uses hard and soft constraints to help allocating heuristic search effort more effectively. The main idea is to modify the objective fitness function by awarding differential incentives according to the defined qualitative preferences, to solution sets which are divided by their satisfaction to constraints. It does not exclude the right to access search spaces that violate some or even all constraints. We test this technique through its application on generating solutions for a classic infinite-horizon extensive-form game. It is solved by an Evolutionary Algorithm incorporated by Incentive method. Experimental results are compared with results from a penalty method and from a non-constraint setting. Statistic analysis suggests that Incentive Method is more effective than the other two techniques for this specific problem.}, bibsource = {DBLP, http://dblp.uni-trier.de}, organisation = {EvoNet}, isbn = {3-540-33143-3}, notes = {Part of \cite{collet:2006:GP} EuroGP'2006 held in conjunction with EvoCOP2006 and EvoWorkshops2006 Two Co-evolving populations. Dividing the cake.}, keywords = {algorithms, genetic programming } } @article{Tsang:2004:DSS, title = {{EDDIE}-Automation, a decision support tool for financial forecasting}, author = {Edward Tsang and Paul Yung and Jin Li}, journal = {Decision Support Systems}, number = {4}, pages = {559--565}, url = {http://www.sciencedirect.com/science/article/B6V8S-4903GV9-1/2/d6ba531a46ce45526ff9015e4447409a}, volume = {37}, year = {2004}, biburl = {http://www.bibsonomy.org/bibtex/23b1ce66f7bc209519bd41c691819b6eb/brazovayeye}, abstract = {Evolutionary Dynamic Data Investment Evaluator (EDDIE) is a genetic programming (GP)-based decision support tool for financial forecasting. EDDIE itself does not replace forecasting experts. It serves to improve the productivity of experts in searching the space of decision trees, with the aim to improve the odds in its user's favour. The efficacy of EDDIE has been reported in the literature. However, discovering patterns in historical data is only the first step towards building a practical financial forecasting tool. Data preparation, rules organisation and application are all important issues. This paper describes an architecture that embeds EDDIE for learning from and monitoring the stock market.}, owner = {wlangdon}, notes = {Special Issue on Data Mining for Financial Decision Making}, keywords = {algorithms, genetic programming } } @incollection{Tsang:2002:gagpcf, title = {{EDDIE} for financial forecasting}, author = {Edward P. K. Tsang and Jin Li}, booktitle = {Genetic Algorithms and Genetic Programming in Computational Finance}, chapter = {7}, editor = {Shu-Heng Chen}, pages = {161--174}, publisher = {Kluwer Academic Press}, url = {http://cswww.essex.ac.uk/CSP/finance/papers/TsangLi-FGP-Chen_CompFinance.pdf}, year = {2002}, biburl = {http://www.bibsonomy.org/bibtex/2c5012f7864619b6c50d278cfab42e8a1/brazovayeye}, abstract = {EDDIE is a genetic-programming based system for channelling expert knowledge into forecasting. FGP-2 is an implementation of EDDIE for Financial forecasting. The novelty of FGP-2 is that, as a forecasting tool, it provides the user with a handle for tuning the precision against the rate of missing opportunities. This allows the user to pick investment opportunities with greater confidence.}, size = {14 pages}, isbn = {0-7923-7601-3}, notes = {part of \cite{chen:2002:gagpcf}}, keywords = {Financial algorithms, decision forecasting, genetic precision, programming, } } @inproceedings{Tsang:2000:COF, title = {Combining Ordinal Financial Predictions with Genetic Programming}, address = {Shatin, N.T., Hong Kong, China}, author = {Edward P. K. Tsang and Jin Li}, booktitle = {Intelligent Data Engineering and Automated Learning - IDEAL 2000: Data Mining, Financial Engineering, and Intelligent Agents}, editor = {Kwong Sak Leung and Lai-Wan Chan and Helen Meng}, month = {13-15 December}, pages = {532--537}, publisher = {Springer-Verlag}, series = {Lecture Notes in Computer Science}, url = {http://link.springer-ny.com/link/service/series/0558/papers/1983/19830532.pdf}, volume = {1983}, year = {2000}, biburl = {http://www.bibsonomy.org/bibtex/21a1b4a9b11d5253dcf02cddb3a242c3e/brazovayeye}, abstract = {Ordinal data play an important part in financial forecasting. For example, advice from expert sources may take the form of {"}bullish{"}, {"}bearish{"} or {"}sluggish{"}, or {"}buy{"} or {"}do not buy{"}. This paper describes an application of using Genetic Programming (GP) to combine investment opinions. The aim is to combine ordinal forecast from different opinion sources in order to make better predictions. We tested our implementation, FGP (Financial Genetic Program-ming), on two data sets. In both cases, FGP generated more accurate rules than the individual input rules.}, issn = {0302-9743}, size = {6 pages}, bibdate = {Tue Sep 10 19:08:58 MDT 2002}, acknowledgement = {}, isbn = {3-540-41450-9}, coden = {LNCSD9}, keywords = {algorithms, genetic programming } } @article{Tsang:2000:JME, title = {{EDDIE} In Financial Decision Making}, author = {Edward P. K. Tsang and Jin Li and Sheri Markose and Hakan Er and Abdel Salhi and Giulia Iori}, journal = {Journal of Management and Economics}, pages = {October}, url = {http://cswww.essex.ac.uk/CSP/finance/papers/EDDIE2000.htm}, year = {2000}, biburl = {http://www.bibsonomy.org/bibtex/2b1b2627f32b1ca5f8c8878aa2077a388/brazovayeye}, abstract = {This paper gives an overview of the EDDIE project. It describes the principles and applications of EDDIE in making financial decisions, including applications to share prices and indices forecasting and arbitrage. EDDIE is designed as an interactive decision tool, not a replacement of expert knowledge. Experts channel their knowledge into the system through (a) selection and preparation of data and (b) providing feedback to EDDIE. EDDIE's main role is to explore interactions between variables and to find thresholds for the variables. Performance of EDDIE depends on both the quality of the users' input and the efficiency of its genetic programming based search engine.}, keywords = {algorithms, financial forecasting genetic programming, } } @article{tsang:1998:eddie, title = {{EDDIE} beats the bookies}, author = {Edward P. K. Tsang and Jin Li and James M. Butler}, journal = {Software: Practice and Experience}, number = {10}, pages = {1033--1043}, url = {http://www3.interscience.wiley.com/cgi-bin/abstract/10007354/START}, volume = {28}, year = {1998}, biburl = {http://www.bibsonomy.org/bibtex/211934e79f1168da4ed289cb04b714a35/brazovayeye}, abstract = {Investment involves the maximisation of return on ones investment whilst minimising risk. Good forecasting, which often requires expert knowledge, can help to reduce risk. In this paper, we propose a genetic programming-based system, EDDIE (Evolutionary Dynamic Data Investment Evaluator), as a forecasting tool. Genetic programming is inspired by evolution theory, and has been demonstrated to be successful in other areas. EDDIE interacts with the users and generates decision trees, which can also be seen as rule sets. We argue that EDDIE is suitable for forecasting because apart from using the power of genetic programming to efficiently search the space of decision trees, it allows expert knowledge to be channelled into forecasting, and it generates rules which can easily be understood and verified. EDDIE has been applied to horse racing and achieved outstanding results. When experimented on 180 handicap races (real data) in the UK, it out-performed other common strategies used in horse race betting by great margins. The idea was then extended to financial forecasting. When tested on historical S&P-500 data EDDIE achieved a respectable annual rate of return over a three and a half year period. While luck may play a part in the success of EDDIE, our experimental results do indicate that EDDIE is a tool which deserves more research. c 1998 John Wiley & Sons, Ltd.}, issn = {0038-0644}, size = {15 pages}, notes = {See also \cite{butler:1995:eddie}}, doi = {doi:10.1002/(SICI)1097-024X(199808)28:10<1033::AID-SPE198>3.0.CO;2-1}, keywords = {algorithms, finance, forecasting, genetic horse investment programming, racing, } } @inproceedings{streeter03, title = {The Root Causes of Code Growth in Genetic Programming}, address = {Essex}, author = {Matthew J. Streeter}, booktitle = {Genetic Programming, Proceedings of EuroGP'2003}, editor = {Conor Ryan and Terence Soule and Maarten Keijzer and Edward Tsang and Riccardo Poli and Ernesto Costa}, month = {14-16 April}, pages = {443--454}, publisher = {Springer-Verlag}, series = {LNCS}, url = {http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=443}, volume = {2610}, year = {2003}, biburl = {http://www.bibsonomy.org/bibtex/20d38d422343e5e7d4fddf13257935f86/brazovayeye}, abstract = {This paper discusses the underlying pressures responsible for code growth in genetic programming, and shows how an understanding of these pressures can be used to use to eliminate code growth while simultaneously improving performance. We begin with a discussion of two distinct components of code growth and the extent to which each component is relevant in practice. We then define the concept of resilience in GP trees, and show that the buildup of resilience is essential for code growth. We present simple modifications to the selection procedures used by GP that eliminate bloat without hurting performance. Finally, we show that eliminating bloat can improve the performance of genetic programming by a factor that increases as the problem is scaled in difficulty.}, organisation = {EvoNet}, publisher_address = {Berlin}, isbn = {3-540-00971-X}, notes = {EuroGP'2003 held in conjunction with EvoWorkshops 2003}, keywords = {algorithms, genetic programming } } @inproceedings{stephenson03, title = {Genetic Programming Applied to Compiler Heuristic Optimization}, address = {Essex}, author = {Mark Stephenson and Una-May O'Reilly and Martin C. Martin and Saman Amarasinghe}, booktitle = {Genetic Programming, Proceedings of EuroGP'2003}, editor = {Conor Ryan and Terence Soule and Maarten Keijzer and Edward Tsang and Riccardo Poli and Ernesto Costa}, month = {14-16 April}, pages = {238--253}, publisher = {Springer-Verlag}, series = {LNCS}, url = {http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=238}, volume = {2610}, year = {2003}, biburl = {http://www.bibsonomy.org/bibtex/272a112dce2af32034dfba6ee4d5dbb39/brazovayeye}, abstract = {Genetic programming (GP) has a natural niche in the optimization of small but high payoff software heuristics. We use GP to optimize the priority functions associated with two well known compiler heuristics: predicated hyperblock formation, and register allocation. Our system achieves impressive speedups over a standard baseline for both problems. For hyperblock selection, application-specific heuristics obtain an average speedup of 23% (up to 73%) for the applications in our suite. By evolving the compiler's heuristic over several benchmarks, the best general-purpose heuristic our system found improves the predication algorithm by an average of 25% on our training set, and 9% on a completely unrelated test set. We also improve a well-studied register allocation heuristic. On average, our system obtains a 6% speedup when it specializes the register allocation algorithm for individual applications. The general-purpose heuristic for register allocation achieves a 3% improvement.}, organisation = {EvoNet}, publisher_address = {Berlin}, isbn = {3-540-00971-X}, notes = {EuroGP'2003 held in conjunction with EvoWorkshops 2003}, keywords = {algorithms, genetic programming } } @inproceedings{smith03, title = {Feature Construction and Selection using Genetic Programming and a Genetic Algorithm}, address = {Essex}, author = {Matthew G. Smith and Larry Bull}, booktitle = {Genetic Programming, Proceedings of EuroGP'2003}, editor = {Conor Ryan and Terence Soule and Maarten Keijzer and Edward Tsang and Riccardo Poli and Ernesto Costa}, month = {14-16 April}, pages = {229--237}, publisher = {Springer-Verlag}, series = {LNCS}, url = {http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=229}, volume = {2610}, year = {2003}, biburl = {http://www.bibsonomy.org/bibtex/2d089752650ac901b05dd0315bed8b9d8/brazovayeye}, abstract = {The use of machine learning techniques to automatically analyse data for information is becoming increasingly widespread. In this paper we examine the use of Genetic Programming and a Genetic Algorithm to pre-process data before it is classified using the C4.5 decision tree learning algorithm. The Genetic Programming is used to construct new features from those available in the data, a potentially significant process for data mining since it gives consideration to hidden relationships between features. The Genetic Algorithm is used to determine which such features are the most predictive. Using ten well-known datasets we show that our approach, in comparison to C4.5 alone, provides marked improvement in a number of cases.}, organisation = {EvoNet}, publisher_address = {Berlin}, isbn = {3-540-00971-X}, notes = {EuroGP'2003 held in conjunction with EvoWorkshops 2003}, keywords = {algorithms, genetic programming } } @inproceedings{sapin03, title = {Research of a cellular automaton simulating logic gates by evolutionary algorithms}, address = {Essex}, author = {Emmanuel Sapin and Olivier Bailleux and Jean-Jacques Chabrier}, booktitle = {Genetic Programming, Proceedings of EuroGP'2003}, editor = {Conor Ryan and Terence Soule and Maarten Keijzer and Edward Tsang and Riccardo Poli and Ernesto Costa}, month = {14-16 April}, pages = {414--423}, publisher = {Springer-Verlag}, series = {LNCS}, url = {http://www.springerlink.com/openurl.asp?genre=article&issn=0302-9743&volume=2610&spage=414}, volume = {2610}, year = {2003}, biburl = {http://www.bibsonomy.org/bibtex/2bf948d6e4c7548b63daf6ce1d20c4c3a/brazovayeye}, abstract = {This paper presents a method of using genetic programming to seek new cellular automata that perform computational tasks. Two genetic algorithms are used : the first one discovers a rule supporting gliders and the second one modifies this rule in such a way that some components appear allowing it to simulate logic gates. The results show that the genetic programming is a promising tool for the search of cellular automata with specific behaviors, and thus can prove to be decisive for discovering new automata supporting universal computation.}, organisation = {EvoNet}, publisher_address = {Berlin}, isbn = {3-540-00971-X}, notes = {EuroGP'2003 held in conjunction with EvoWorkshops 2003}, keywords = {algorithms, genetic programming } }