Wellbore Trajectory Generation With Genetic Algorithm‐Assisted Generative Adversarial Network

ABSTRACT The generation of three‐dimensional double‐build wellbore trajectories for horizontal wells is a critical task in petroleum engineering, constrained by complex underground obstacles and trajectory target requirements. Existing frameworks rely on random exploration for new solutions, unable to learn from the fitness landscape, thus failing to proactively move toward high‐quality regions. Consequently, many generated solutions are discarded as subpar, significantly raising computational costs from invalid attempts. Neural networks, while capable of learning solution distributions, lack targeted local search and easily get trapped in local optima. To address these issues, this paper proposes a novel framework named GA‐optiGAN, which integrates the Generative Adversarial Network and Genetic Algorithm. First, we construct a dynamically maintained candidate solution set, where an exponential decay contraction factor is adopted for dynamic scale adjustment. This encourages our framework to focus on high‐quality regions in the later stage while maintaining a consistent exploration‐exploitation balance throughout the process. Meanwhile, Genetic Algorithm performs parameter‐level fine‐tuning within this set, which allows for refined local search and prevents entrapment in local optima. Second, we design a tanh‐based range scaling module, embedding the constraint ranges of wellbore trajectory parameters into GA‐optiGAN's generator to ensure engineering‐compliant parameter outputs and avoid invalid solutions. Third, we design an objective function centered on minimizing trajectory length, and enforce critical constraints via a two‐stage penalty mechanism that eliminates solutions failing to satisfy these constraints, including obstacle avoidance, target‐hitting accuracy, and parametric compatibility. Experimental results demonstrate that GA‐optiGAN achieves superior performance over state‐of‐the‐art frameworks in generating three‐dimensional double‐build wellbore trajectories.

Authors

Institutions

Publication Details

Journal
Concurrency and Computation Practice and Experience
Published
2026-09-27
DOI
https://doi.org/10.1002/cpe.70972
Primary Topic
Reservoir Engineering and Simulation Methods
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Wellbore Trajectory Generation With Genetic Algorithm‐Assisted Generative Adversarial Network

Haitao Zhang, Mingao Li
Concurrency and Computation Practice and Experience
Reservoir Engineering and Simulation Methods
article

Wellbore Trajectory Generation With Genetic Algorithm‐Assisted Generative Adversarial Network

Haitao Zhang, Mingao Li
article en

Abstract

ABSTRACT The generation of three‐dimensional double‐build wellbore trajectories for horizontal wells is a critical task in petroleum engineering, constrained by complex underground obstacles and trajectory target requirements. Existing frameworks rely on random exploration for new solutions, unable to learn from the fitness landscape, thus failing to proactively move toward high‐quality regions. Consequently, many generated solutions are discarded as subpar, significantly raising computational costs from invalid attempts. Neural networks, while capable of learning solution distributions, lack targeted local search and easily get trapped in local optima. To address these issues, this paper proposes a novel framework named GA‐optiGAN, which integrates the Generative Adversarial Network and Genetic Algorithm. First, we construct a dynamically maintained candidate solution set, where an exponential decay contraction factor is adopted for dynamic scale adjustment. This encourages our framework to focus on high‐quality regions in the later stage while maintaining a consistent exploration‐exploitation balance throughout the process. Meanwhile, Genetic Algorithm performs parameter‐level fine‐tuning within this set, which allows for refined local search and prevents entrapment in local optima. Second, we design a tanh‐based range scaling module, embedding the constraint ranges of wellbore trajectory parameters into GA‐optiGAN's generator to ensure engineering‐compliant parameter outputs and avoid invalid solutions. Third, we design an objective function centered on minimizing trajectory length, and enforce critical constraints via a two‐stage penalty mechanism that eliminates solutions failing to satisfy these constraints, including obstacle avoidance, target‐hitting accuracy, and parametric compatibility. Experimental results demonstrate that GA‐optiGAN achieves superior performance over state‐of‐the‐art frameworks in generating three‐dimensional double‐build wellbore trajectories.

Concurrency and Computation Practice and ExperienceVol. 38(19)
Beijing University of Posts and Telecommunications (CN), State Key Laboratory of Networking and Switching Technology
Openalex Percentile: Top 16%
Reservoir Engineering and Simulation Methods
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.