Multi-criteria optimization of planning and operation of irrigation systems including the local rainfall characteristics with a special focus on sustainability

目前的灌溉实践一般只着眼于优化一个单一目标，即短期经济产量。因此，在寻找最适合这一目标的灌溉管理战略时，不包括可持续性方面或其他重要标准。同时，气象边界条件的随机性和田间土壤水力参数的变异性的影响通常被忽略。为了解决这些限制所带来的问题，我们提出了一个更全面的灌溉管理优化的观点。我们计划组装一个模拟系统，它可以模拟所有相关的灌溉相关过程的考虑位置。在此基础上，我们将开发的优化程序，不仅可以处理随机边界条件，但也多维目标标准，实现整体最优解。我们设想推广这种解决方案的基础上，一个扩展的土壤转移的方法。这种方法适用于周围的领域，表现出灌溉相关的土壤水分范围内-类似的土壤水力特性。该工具采用神经网络和基于实例的推理技术，操作简单，能够满足灌溉实践的需要。

Current irrigation practice generally aims at optimising one single goal only, namely, the short term economical yield. Thus, sustainability aspects or other important criteria are not included when looking for the irrigation management strategies best suited to this goal. As well, the influence of the stochastic properties of meteorological boundary conditions and the variability of soil hydraulic parameters within a field are commonly neglected. To tackle the problems which arise from these limitations, we propose a more comprehensive view of irrigation management optimisation. We plan to assemble a simulation system which allows simulating all relevant irrigation related processes for a considered location. On this basis we will develop optimization procedures that can handle not only stochastic boundary conditions but also multidimensional target criteria for achieving an overall optimal solution. We envisage generalising this solution on the basis of an extended pedo-transfer approach. This approach applies to surrounding fields, which exhibit within an irrigation relevant soil moisture range - similar soil hydraulic properties. Using neural networks and case based reasoning techniques, the new tool will finally satisfy the requirements of irrigation practice for a convenient and straightforward application.