Speed Benchmarking of Genetic Programming Frameworks

TL;DR

This paper benchmarks various genetic programming frameworks, highlighting TensorGP's significant speed improvements through vectorization and TensorFlow integration, enabling larger problem domains and more accurate solutions.

Contribution

Introduces TensorGP, a new Python-based GP framework leveraging TensorFlow for accelerated evaluation, and provides comprehensive performance benchmarks comparing it to existing approaches.

Findings

TensorGP achieves over 100x speedup on large problems

Vectorization significantly improves GP evaluation times

Larger domain evaluations lead to more accurate solutions

Abstract

Genetic Programming (GP) is known to suffer from the burden of being computationally expensive by design. While, over the years, many techniques have been developed to mitigate this issue, data vectorization, in particular, is arguably still the most attractive strategy due to the parallel nature of GP. In this work, we employ a series of benchmarks meant to compare both the performance and evolution capabilities of different vectorized and iterative implementation approaches across several existing frameworks. Namely, TensorGP, a novel open-source engine written in Python, is shown to greatly benefit from the TensorFlow library to accelerate the domain evaluation phase in GP. The presented performance benchmarks demonstrate that the TensorGP engine manages to pull ahead, with relative speedups above two orders of magnitude for problems with a higher number of fitness cases. Additionally, as a consequence of being able to compute larger domains, we argue that TensorGP performance gains aid the discovery of more accurate candidate solutions.