We report results from a series of studies coevolving players for simple Rock–Paper–Scissors game... more We report results from a series of studies coevolving players for simple Rock–Paper–Scissors games. These results demonstrate that Current Individual versus Ancestral Opponent (CIAO) plots, which have been proposed as a visualization technique for detecting both coevolutionary progress and coevolutionary cycling, suffer from ambiguity with respect to an important but rarely discussed class of cyclic behavior. While regular cycling manifests itself as a characteristic banded plot, irregular cycling produces an irregular tartan pattern which is also consistent with random drift through strategy space. Although this tartan pattern is often reported in the literature on coevolutionary algorithms, it has received little attention or analysis. Here we argue that irregular cycling will tend to be more prevalent than regular cycling, and that it corresponds to a class of coevolutionary scenario that is important both theoretically and in practice. As such, it is desirable that we improve our ability to distinguish its occurrence from that of random drift, and other forms of coevolutionary dynamic.
Currently, evolutionary computation (EC) typically takes place in batch mode: algorithms are run ... more Currently, evolutionary computation (EC) typically takes place in batch mode: algorithms are run au- tonomously, with the user providing little or no interven- tion or guidance. Although it is rarely possible to specify in advance, on the basis of EC theory, the optimal evo- lutionary algorithm for a particular problem, it seems likely that experienced EC practitioners possess consid- erable
We report results from a series of studies coevolving players for simple Rock–Paper–Scissors game... more We report results from a series of studies coevolving players for simple Rock–Paper–Scissors games. These results demonstrate that Current Individual versus Ancestral Opponent (CIAO) plots, which have been proposed as a visualization technique for detecting both coevolutionary progress and coevolutionary cycling, suffer from ambiguity with respect to an important but rarely discussed class of cyclic behavior. While regular cycling manifests itself as a characteristic banded plot, irregular cycling produces an irregular tartan pattern which is also consistent with random drift through strategy space. Although this tartan pattern is often reported in the literature on coevolutionary algorithms, it has received little attention or analysis. Here we argue that irregular cycling will tend to be more prevalent than regular cycling, and that it corresponds to a class of coevolutionary scenario that is important both theoretically and in practice. As such, it is desirable that we improve our ability to distinguish its occurrence from that of random drift, and other forms of coevolutionary dynamic.
We report results from a series of studies coevolving players for simple Rock–Paper–Scissors game... more We report results from a series of studies coevolving players for simple Rock–Paper–Scissors games. These results demonstrate that Current Individual versus Ancestral Opponent (CIAO) plots, which have been proposed as a visualization technique for detecting both coevolutionary progress and coevolutionary cycling, suffer from ambiguity with respect to an important but rarely discussed class of cyclic behavior. While regular cycling manifests itself as a characteristic banded plot, irregular cycling produces an irregular tartan pattern which is also consistent with random drift through strategy space. Although this tartan pattern is often reported in the literature on coevolutionary algorithms, it has received little attention or analysis. Here we argue that irregular cycling will tend to be more prevalent than regular cycling, and that it corresponds to a class of coevolutionary scenario that is important both theoretically and in practice. As such, it is desirable that we improve our ability to distinguish its occurrence from that of random drift, and other forms of coevolutionary dynamic.
Currently, evolutionary computation (EC) typically takes place in batch mode: algorithms are run ... more Currently, evolutionary computation (EC) typically takes place in batch mode: algorithms are run au- tonomously, with the user providing little or no interven- tion or guidance. Although it is rarely possible to specify in advance, on the basis of EC theory, the optimal evo- lutionary algorithm for a particular problem, it seems likely that experienced EC practitioners possess consid- erable
We report results from a series of studies coevolving players for simple Rock–Paper–Scissors game... more We report results from a series of studies coevolving players for simple Rock–Paper–Scissors games. These results demonstrate that Current Individual versus Ancestral Opponent (CIAO) plots, which have been proposed as a visualization technique for detecting both coevolutionary progress and coevolutionary cycling, suffer from ambiguity with respect to an important but rarely discussed class of cyclic behavior. While regular cycling manifests itself as a characteristic banded plot, irregular cycling produces an irregular tartan pattern which is also consistent with random drift through strategy space. Although this tartan pattern is often reported in the literature on coevolutionary algorithms, it has received little attention or analysis. Here we argue that irregular cycling will tend to be more prevalent than regular cycling, and that it corresponds to a class of coevolutionary scenario that is important both theoretically and in practice. As such, it is desirable that we improve our ability to distinguish its occurrence from that of random drift, and other forms of coevolutionary dynamic.
Position paper introducing an AI research methodology for investigating dynamics and stability of... more Position paper introducing an AI research methodology for investigating dynamics and stability of financial markets. Controlled experiments using an experimental financial marketplace containing humans and adap- tive trading agents are performed; with brain activity of hu- man participants captured using an EEG headset. Machine learning is applied to brain and trading data to develop an adaptive multi-agent model of markets that can be optimised for stability using a co-evolutionary GA framework. Specif- ically, financial market circuit breakers—mechanisms for limiting or halting trading on an exchange—are addressed; a pertinent real-world problem of trans-national importance. Ex post circuit breakers that halt a market after a sudden price swing will be analysed and optimised. Further, ex ante circuit breakers, which are triggered before problems emerge, will be investigated and explored. The development of reliable “anti-crash” trading technology would have globally significant economic, social, and political impact.
As we build increasingly large scale systems (and systems of systems), the level of complexity is... more As we build increasingly large scale systems (and systems of systems), the level of complexity is also rising. We still expect people to intervene when things go wrong, however, and to diagnose and fix the problems. Aviation has a history of developing systems with a very good safety record. Domains such as high frequency trading (HFT), however, have a much more chequered history. We note that there are several parallels that can be drawn between aviation and HFT. We highlight the ironies of automation that apply to HFT, before going on to identify several lessons that have been used to improve safety in aviation and show how they can be applied to increase the resilience of HFT in particular.
Uploads