GPTP 2009: An Example of Evolvability

Genetic programming theory and practice: a fifteen-year trajectory

“…[14] GPTP VII: 2009. O'Reilly et al [5] identified a number of GP challenges, some of which we have encountered before: economic resource usage (processor, memory, disk), ensuring quality results that are usable (e.g., humaninterpretable, trustworthy, or predictive on very different inputs), reliable convergence, and addressing the wide range of problem domains (representations, operators, and so forth).…”

“…Practitioners showcased time-series modeling, high-dimensional symbolic regression and classification, and financial applications. "The consensus among the participants this year," wrote O'Reilly et al [5], "was that genetic programming has reached a watershed in terms of practicality for a well-defined range of applications." This is not quite straightforward because it necessitates an "appropriate determination of algorithm techniques, representation, operators, and fitness function."…”

“…This is not quite straightforward because it necessitates an "appropriate determination of algorithm techniques, representation, operators, and fitness function." [5] The GPTP events were seen to promote successful techniques for alleviating commonly occurring problems such as premature convergence, bloat, and scalability. The focus on best-practice approaches presented a new direction for GP.…”