چكيده به لاتين
Managing water resources to meet the growing demands for energy and food while preserving the environment, is a significant challenge of our time, especially when dealing with the complex natural-human system exposed to various uncertainties in resources. Nowadays, almost all major watersheds worldwide can be considered as a natural-human system where heterogeneous human activities impact and are impacted by the natural hydrological cycle. Agent-based modeling and simulation offer a novel approach to model systems consisting of independent and interacting agents. The goal of this research is to examine human decisions under environmental and human uncertainties by defining a simulation-optimization model, understanding how farmers' behavior patterns and crop cultivation change affect the collective profit of the system and the regional cropping pattern under different risk levels and irrigation strategies. Furthermore, this model is made available as a flexible web-based tool to enhance user collaboration and increase user awareness of their water resource decisions. Human decision-making processes in the agent-based model are outlined and implemented using a stochastic optimization model in the Python environment. This optimization model has a non-deterministic objective function for maximizing the economic gross profit of agents and several constraints, which depend on multiple input parameters. These parameters include:1) Environmental uncertainty - precipitation: A precipitation-runoff regression model is considered in the model to account for this parameter, simulating runoff scenarios for the Maroun and Jirahi rivers. This uncertainty is incorporated into a calibrated WEAP (Water evaluation And Planning) water resource simulation model to estimate available water resources for farmers, and the model is automated in the MATLAB environment), 2) Human uncertainty - crop price growth rates: A probability distribution function is determined, and samples from it are incorporated into the objective function. And 3) Crop performance and water consumption: Soil moisture dynamic equations for the region's suitable soil types for crop cultivation, considering traditional and industrial irrigation practices as well as standard and deficit irrigation strategies, are solved to calculate these parameters. The problem is solved using a stochastic scenario-based approach, and the Expected Gain Confidence Limit method is used to evaluate the stochastic objective function, which involves comparing the means and variances of two scenarios and the risk parameter (λ). From a practical standpoint, many environmental-human models have been developed for surface waters but are often inaccessible to non-developer users. By utilizing web technologies, designing a web-based application from environmental-human models transforms into a user-friendly process, enhancing data handling capabilities, model accessibility and service collaboration. The proposed framework of this study was tested in the Maroun-Jirahi watershed. Based on the results of this model, it predicts that farmers with a high risk-taking attitude (λ 0). Will have. In modern irrigation, under the stress of low irrigation, for an agent with a risk-taking attitude, 23% more profit was obtained than for a neutral agent and 53% for a risk-averse agent. Also, in other types and strategies of irrigation, the amount of profit of the risk-taking agent to the neutral and risk-averse agent was on average equal to 17% and 30%. In the whole system, it can be seen that the system is more inclined towards the low irrigation strategy. In the scenario of leaving the agents free to determine the minimum and maximum area under cultivation for each product, the risk-taking agents changed their cultivation pattern and as a result increased profits, and the risk-averse farmers preferred to have a more uniform cultivation pattern to reduce the amount of existing risks and the result was no change or a decrease in the profit of the system.
