INTERACTIVE EVOLUTION FOR SYSTEMATIC EXPLORATION OF A PARAMETER SPACE

Interactive evaluation of a fitness function for a genetic algorithm through direct manipulation is known as interactive evolution. Because it removes the need to specify a fitness function prior to exploration, the user can change evaluation criteria over time. This is especially important when the fitness function is unknown or not easily specified beforehand. This paper describes a system that allows the user to evaluate and evolve small collections of candidate solutions that represent the range of available solutions in order to explore parameter ranges of interest. Although users may be expert with respect to their particular tasks, some may be novices in relation to their software. For them, interactive evolution removes syntactic barriers to the specification of alternative candidate solutions and navigation amongst them. Expert users benefit from exposure to candidate solutions outside of their prior experience and so may find solutions that better meet their needs. INTRODUCTION Consider that every parameter under consideration has some permissible values, either discrete or continuous. The space defined by these parameters is the Cartesian product of all values for all parameters. For N parameters, each N-tuple of values is a distinct point in the parameter space, as follows: <v1,v2,..., vN> Œ P1 x P2 x ... x PN One might easily encounter what is called the “curse of dimensionality” or a “combinatorial explosion” when exploring this space. If each of 3 parameters have just 10 values, there are 1000 distinct tuples in this space, making an exhaustive search overwhelming. It is also fruitful to consider this parameter space as a solution space (Simon, 1977), since the expectation exists that a solution to the current problem can be found amongst the combinations of parameters values. The computer has the promise to democratize exploration of these parameter spaces, as suggested by Jessup (1992), for example. The promise, however, will not be fully realized until software addresses the issue of access. In general, the relationship between humans and computers can be viewed as either automatic, manual, or augmented (Kochhar et al., 1991). Automatic systems, like Mackinlay’s APT [A Presentation Tool] (1986), are generally prescriptive and work with very little input from the user to determine the visual representation for the user. Manual systems, like AVS [Advanced Visualization