Optimized Bidirectional Gas–water Displacement in High-inclination Reservoirs Using an Improved Geyser-inspired Algorithm

A field-scale high-inclination well pattern model is constructed, and an improved Geyser-inspired optimization algorithm (IGEA) is proposed by integrating Halton sequence initialization, tangent flight strategy, and sine–cosine perturbation mechanism to optimize key parameters such as injection/production rates, cycles, and switching timing. Results show that IGEA outperforms GEA and PSO in terms of convergence speed, optimization accuracy, and stability, with average net present value improved by 1% and 6%, respectively. Further analysis indicates that different injection strategies have a significant impact on development performance: water-alternating-gas injection performs best in terms of both net present value and oil recovery, followed by continuous gas injection, while intermittent gas injection is less effective. The choice of objective function also affects optimization results: NPV-oriented strategies help to improve early production and capital recovery, whereas production-oriented strategies place greater emphasis on maintaining injection–production balance and maximizing ultimate recovery. This study provides an effective optimization approach for the development of high-inclination reservoirs.