Abstract
As a clean backup energy source, natural gas hydrates have garnered significant global attention, making it crucial to establish models for predicting the total volume of regional resources. This article employs the volumetric method as the foundation for predictions, utilizing data from 14 wells within the study area as test dataset. Initially, we choose the KNN interpolation algorithm to establish the nearest neighbor value, subsequently estimating the effective area and delineating the distribution range of hydrate resources. Subsequently, using the Kolmogorov-Smirnov test (KS test), we discover that the parameters of natural gas hydrate resources approximately adhere to both the Weber and Poisson distributions, depending on the coordinates. And we delineate the general distribution pattern of natural gas hydrates and estimate the resource quantity within each well as well as the total regional volume. Lastly, employing a combination of local search algorithms—greedy, simulated annealing, and Monte Carlo—along with a global search algorithm (genetic algorithm), we predict the location of wells with the greatest potential future resource volume. After thorough consideration, we identify five potential well locations and assess the applicability of our current algorithmic model in various environments, along with its future prospects.