1,721,243 research outputs found
Adaptive business intelligence: Three case studies
No full textZbigniew Michalewicz, Martin Schmidt, Matthew Michalewicz and Constantin Chiria
Parameter control in evolutionary algorithms
Full text linkThe issue of setting the values of various parameters of an evolutionary algorithm is crucial for good performance. In this paper we discuss how to do this, beginning with the issue of whether these values are best set in advance or are best changed during evolution. We provide a classification of different approaches based on a number of complementary features, and pay special attention to setting parameters on-the-fly. This has the potential of adjusting the algorithm to the problem while solving the problem. This paper is intended to present a survey rather than a set of prescriptive details for implementing an EA for a particular type of problem. For this reason we have chosen to interleave a number of examples throughout the text. Thus we hope to both clarify the points we wish to raise as we present them, and also to give the reader a feel for some of the many possibilities available for controlling different parameters. © Springer-Verlag Berlin Heidelberg 2007.A. E. Eiben, Zbigniew Michalewicz, M. Schoenauer and J. E. Smit
Interpretable multi-criteria fuzzy rule based decision models for hedge fund management
No full textThis paper describes an approach to constructing fuzzy rules for predictive modeling that involves a local search heuristic and an evolutionary algorithm. This approach is applied for learning strategies to manage a portfolio that comprises positions in the share market. We provide experimental results comparing the approach to random strategies and the market index. A non-linear prediction model that relates asset performance to a large set of explanatory variables is represented with fuzzy rules. Rulebases are combined to build multi-criteria recommendations for trading decisions that consider different forecast horizons and both risk and return criteria.Adam Ghandar, Zbigniew Michalewicz and Ralf Zurbrueg
Enhancing profitability through interpretability in algorithmic trading with a multiobjective evolutionary fuzzy system
No full textThis paper examines the interaction of decision model complexity and utility in a computational intelligence system for algorithmic trading. An empirical analysis is undertaken which makes use of recent developments in multiobjective evolutionary fuzzy systems (MOEFS) to produce and evaluate a Pareto set of rulebases that balance conflicting criteria. This results in strong evidence that controlling portfolio risk and return in this and other similar methodologies by selecting for interpretability is feasible. Furthermore, while investigating these properties we contribute to a growing body of evidence that stochastic systems based on natural computing techniques can deliver results that outperform the market.Adam Ghandar, Zbigniew Michalewicz and Ralf Zurbrueg
Short and long term memory in coevolution
No full textGames provide the perfect test bed for measuring the effectiveness ofcomputer generated strategies in a competitive and fun environment.Over the years many different games have been tackled by researchersof computational intelligence with the purpose of creating an intelligentcomputer player that can challenge human players. In this paper the au-thors summarize the research performed over the past two years basedon the game of Tempo, where the coevolved strategies were representedby logic rule bases with an adaptive memory. The experiments were setup to investigate the effectiveness of various memory structures in a co-evolutionary game system, as well as the effectiveness of various recallprocesses.Phillipa Avery, Zbigniew Michalewicz and Martin Schmid
Evaluation of Intelligent Quantitative Hedge Fund Management
No full textThis paper examines an intelligent recommendation strategy implementation for managing a long short hedge fund and reports on performance during market conditions at the onset of the liquidity crisis. A hedge fund utilizes long and short trading to manage an investment portfolio consisting of allocations to cash and share equity positions. This results in a combined long short portfolio that is leveraged to obtain a potentially greater market exposure with borrowed cash from short selling and is also hedged to protect against market downturns. The paper also examines effects of parameters for fuzzy rule base specification on trading performance.Muneer Buckley, Adam Ghandar, Zbigniew Michalewicz, Ralf Zurbrueg
Return performance volatility and adaptation in an automated technical analysis approach to portfolio management
No full textCopyright © 2009 John Wiley & Sons, Ltd.AbstractThis paper discusses the design of a quantitative computational intelligence portfolio management system and evaluates the advantages of some adaptive mechanisms to enable the system to adjust its management approach as market conditions change. A detailed analysis of the performance of the system outside is also provided. It is found that an adaptive methodology where trading rules are able to adjust to market conditions performs better, having greater excess returns and lower volatility than a fixed rule approach. We consider several performance metrics, including portfolio alpha and information content. Copyright © 2009 John Wiley & Sons, Ltd.Adam Ghandar, Zbigniew Michalewicz and Ralf Zurbrueg
Evolving fuzzy rules: evaluation of a new approach
No full textEvolutionary algorithms have been successfully applied to optimize the rulebase of fuzzy systems. This has lead to powerful automated systems for financial applications. We experimentally evaluate the approach of learning fuzzy rules by evolutionary algorithms proposed by Kroeske et al. [10]. The results presented in this paper show that the optimization of fuzzy rules may be universally simplified regardless of the complex fitness surface for the overall optimization process. We incorporate a local search procedure that makes use of these theoretical results into an evolutionary algorithms for rule-base optimization. Our experimental results show that this improves a state of the art approach for financial applications. © 2010 Springer-Verlag.Adam Ghandar, Zbigniew Michalewicz and Frank Neumannhttp://portal.acm.org/citation.cfm?id=194748
Intelligent decision support: A fuzzy stock ranking system
No full textThis paper presents an intelligent decision support system for financial portfolio management. An adaptive business intelligence approach combines optimization, forecasting and adaptation with application specific financial information processing and quantitative investment paradigms. The methodology involves constructing a ranking of stocks by strength of a buy or sell recommendation which is inferred using an adapting forecasting model that considers a range of factors. These include company balance sheet information, market price and trading volume as well as the wider economy. The system adjusts its prediction model dynamically as market conditions change. An evolving fuzzy rule base mechanism encodes a model of relationships between model factors and a recommendation to buy, sell or hold securities.Adam Ghandar, Zbigniew Michalewicz and Ralf Zurbrueg
- …
