Order-Aware GSGP in Active Factor Allocation

Created by W.Langdon from gp-bibliography.bib Revision:1.9225

@InProceedings{DBLP:conf/ppsn/BunjerdtaweepornM26,

• author = "Kritpol Bunjerdtaweeporn and Alberto Moraglio",
• editor = "Giovanni Iacca and Giorgia Nadizar and Anil Yaman and Doina Bucur and Antonio Della Cioppa and Ting Hu and Eric Medvet and Sarah L. Thomson",
• title = "Order-Aware {GSGP} in Active Factor Allocation",
• booktitle = "Parallel Problem Solving from Nature - {PPSN} {XIX} - 19th International Conference, {PPSN} 2026, Trento, Italy, August 29 - September 2, 2026, Proceedings, Part {IV}",
• series = "Lecture Notes in Computer Science",
• volume = "16988",
• pages = "168--183",
• publisher = "Springer",
• year = "2026",
• keywords = "genetic algorithms, genetic programming",
• URL = "https://doi.org/10.1007/978-3-032-36229-2_11",
• DOI = "10.1007/978-3-032-36229-2_11",
• timestamp = "Wed, 09 Sep 2026 18:09:11 +0200",
• biburl = "https://dblp.org/rec/conf/ppsn/BunjerdtaweepornM26.bib",
• bibsource = "dblp computer science bibliography, https://dblp.org",
• abstract = "A recent extension of geometric semantic genetic programming (GSGP) is order-aware GSGP. This approach employs geometric semantic operators for real-valued functions to directly search for functions that match the target output ordering behaviour using the Kemeny distance metric, while maintaining a unimodal fitness landscape. Factor timing is an active investment strategy that adjusts portfolio exposure to risk factors expected to outperform. Since the task is fundamentally about ranking factors, order-aware GSGP is a natural fit. We evaluate the efficacy of order-aware GSGP in this context. Our empirical results show that order-aware GSGP significantly outperforms standard GP with Kemeny distance as a fitness function, suggesting a critical role for the unimodality of the fitness landscape in order semantic space. Portfolios constructed using order-aware GSGP achieve higher mean returns and lower turnover than those built with traditional GSGP. Strategies derived from order-aware GSGP also outperform a static equal-weight benchmark that most active investment approaches fail to beat, even after accounting for transaction costs.",
• notes = "https://ppsn2026.disi.unitn.it/",

}

Genetic Programming entries for Kritpol Bunjerdtaweeporn Alberto Moraglio

Citations