Dave Cliff

Evolution of market mechanism through a continuous space of auction-types

A continuous space of auction mechanisms is explored via a genetic algorithm, with ZIP artificial trading agents ope rating in the evolved markets. The space of possible auction-types includes the Continuous Double Auction and also two purely one-sided mechanisms, yet hybrids of these auction types are regularly found to give the most desirable market dynamics.

Evolving market design in zero-intelligence trader markets

The Continuous Double Auction (CDA) is one of the most popular of market-mechanisms. Previous work using genetic algorithms (GAs) for automated mechanism design by Cliff has shown that previously unexplored hybrid variants of the traditional CDA can lead to preferable market dynamics. Cliff s results were based on experiments conducted using a computational simulation of the CDA populated by electronic Zero Intelligence Plus (ZIP) traders and his work uses a GA to co-evolve the market mechanism with the ZIP agent parameters. We use a self-adaptive evolutionary strategy (ES) to explore the space of possible auction types in a CDA populated by Gode and Sunders cognitively simple Zero Intelligence Constrained (ZI-C) traders. We show that non-standard CDAs are still preferred over traditional variants and our results provide the first demonstration that non-standard variants of the CDA can provide favorable dynamics for trading strategies other than ZIP.

Evolved hybrid auction mechanisms in non-ZIP trader marketplaces

A previous paper by D. Cliff (see ibid., 2002) demonstrated that a genetic algorithm could be used to automatically discover new optimal auction mechanisms for automated electronic marketplaces populated by software-agent traders. Significantly, the new auction mechanisms are often unlike traditional mechanisms designed by humans for human traders; rather, they are peculiar hybrid mixtures of established styles of mechanism. This previous work used software agents running the ZIP trader algorithm (recently shown to outperform human traders). We provide the first demonstration that qualitatively similar results (i.e., non-standard hybrid mechanism designs being optimal) are also given when similar experiments are performed using a different trader algorithm, namely Gode & Sunders (1993) ZI-C traders. Thus, the paper is the first to offer significant evidence that evolved hybrid auction mechanisms may be found that out-perform traditional market mechanisms for many styles of trader-agent.

hpDJ: An automated DJ with floorshow feedback

The hpDJ system described here goes some way towards replacing the tasks performed by human DJs. It has potential use as a component in the user-interface to audio-based consumer digital entertainment systems, converting the audio data stored on such systems from a set of songs into a continuous seamless mix. Such mixes are suitable for play-out over streaming media (e.g., in personalized internet radio), or for writing to an appropriate recording medium (such as CD, the hard disk of an iPod, or a flash ROM card) for subsequent playback, or for playing to crowds of dancers in real nightclubs. Results from the nightclub experiment are promising, and our subsequent development of monitoring technology allows crowd feedback to influence hpDJ’s choices of songs, making it even more human-like. The use of human-inspired heuristics in dynamically selecting customized DSP filters for the cross-fade has the potential to allow hpDJ to perform cross-fades in ways that would be virtually impossible for a human DJ playing live. While there is a growing market for software products that give a “virtual” version of traditional human-DJ hardware, and while MixMeister provides a pleasant interface to a set of software tools that allow an unskilled human to create professional-quality continuous mixes, hpDJ as described here is as far as we know the first and only system that aims to totally automate the tasks performed by a human nightclub DJ, including dynamically reacting to the responses from the crowd in real-time. Although we have yet to test Version 2 in a real nightclub, it is clear that the prospect of crowd monitoring opens up new possibilities for the computer-assisted composition of music. But, whereas most computer-aided music composition systems assume a single human author working with the machine, the vision in hpDJ is that the author is an entire crowd of participants, collaborating indirectly, giving feedback as they consume the music. That feedback being generated either actively by the members of the crowd hitting the buttons on their voting watches; or passively by them merely dancing and having a good time, while the computer watches them.

Using a Genetic Algorithm to Design and Improve Storage Area Network Architectures

Designing storage area networks is an NP-hard problem. Previous work has focused on traditional algorithmic techniques to automatically determine fabric requirements, network topology, and flow routes. This paper presents work performed with a genetic algorithm to both improve designs developed with heuristic techniques and to create new designs. For some small networks (10 hosts, 10 devices, and single-layered) we find that we can create networks which result in savings of several thousand dollars over previously established methods. This paper is the first publication, to our knowledge, to describe the successful application of this technique to storage area network design.

An Ant-Inspired Technique for Storage Area Network Design

Designing storage area networks is an NP-hard problem. Previous work has focused on traditional algorithmic techniques to automatically determine fabric requirements, network topology, and flow routes. This paper looks at the ability of an ant colony optimisation algorithm to evolve new architectures. For some small networks (10 hosts, 10 devices, and single-layered) we find that we can create networks which result in savings of several thousand dollars over previously established methods. This paper is the first publication, to our knowledge, to describe the successful application of this technique to storage area network design.

Evolutionary optimization of ZIP60: a controlled explosion in hyperspace

The ZIP adaptive trading algorithm has been demonstrated to outperform human traders in experimental studies of continuous double auction (CDA) markets. The original ZIP algorithm requires the values of eight control parameters to be set correctly. A new extension of the ZIP algorithm, called ZIP60, requires the values of 60 parameters to be set correctly. ZIP60 is shown here to produce significantly better results than the original ZIP (called ZIP8 hereafter). A genetic algorithm (GA) is used to search the 60-dimensional ZIP60 parameter space, and it finds parameter vectors that yield ZIP60 traders with mean scores significantly better than those of ZIP8s. This paper shows that this optimizing evolutionary search works best when the GA itself controls the dimensionality of the search-space, so that the search commences in an 8-d space and thereafter the dimensionality of the search-space is gradually increased by the GA until it is exploring a 60-d space. Furthermore, the results from ZIP60 cast some doubt on prior ZIP8 results concerning the evolution of new hybrid auction mechanisms that appeared to be better than the CDA.

Making money from agents

The discipline of agent technology is becoming a mature research area, both from a scientific and an engineering perspective. Strong theoretical and empirical analysis, together with practical experience of deploying systems to solve real problems, give this community the opportunity to make an impact in the world of business. At UKMAS98, a panel of academics and industrialists met to discuss exactly how this could be done; how could agents and multi-agent systems be used to actually make money?