Reasoning about the factors underlying habitat connectivity and the inter-habitat movement of species is essential to many areas of biological inquiry. In order to better describe and understand the ways in which the landscape may support species movement, an increasing amount of research has focused on identification of paths or corridors that may be important in providing connectivity among habitat. The least-cost path problem has proven to be an instrumental analytical tool in this sense. A complicating aspect of such path identification methods is how to best reconcile and integrate the array of criteria or objectives that species may consider in traversal of a landscape. In cases where habitat connectivity is thought to be influenced or guided by multiple objectives, numerous solutions to least-cost path problems can exist, representing tradeoffs between the objectives. In practice though, identification of these solutions can be very challenging and as such, only a small proportion of them are typically examined leading to a weak characterization of habitat connectivity. To address this computational challenge, a multiobjective optimization framework is proposed. A generalizable multiobjective least-cost path model is first detailed. A non-inferior set estimation (MONISE) algorithm for identifying supported efficient solutions to the multiobjective least-cost path model is then described. However, it is well known that unsupported efficient solutions (which are equally important) can also exist, but are typically ignored given that they are more difficult to identify. Thus, to enable the identification of the full set of efficient solutions (supported and unsupported) to the multiobjective model, a multi-criteria labeling algorithm is then proposed. The developed framework is applied to assess different conceptualizations of habitat connectivity supporting amphibian movement in a wetland system. The results highlight the range of tradeoffs in characterizations of connectivity that can exist when multiple objectives are thought to contribute to movement decisions and that the number of unsupported efficient solutions (which are typically ignored) can vastly outweigh that of the supported efficient solutions.

对栖息地连通性和物种在栖息地间移动的基础因素进行推理对于许多生物学探究领域都是至关重要的。为了更好地描述和理解景观支持物种运动的方式，越来越多的研究集中在识别路径或走廊上，这些路径或走廊可能对提供栖息地之间的连通性很重要。从这个意义上说，成本最低的路径问题已被证明是一种工具分析工具。这种路径识别方法的一个复杂方面是如何最好地协调和整合物种在穿越景观时可能考虑的一系列标准或目标。如果认为栖息地的连通性受到多个目标的影响或指导，那么可以找到许多解决成本最低的路径问题的方法，代表目标之间的权衡。但是在实践中，识别这些解决方案可能会非常具有挑战性，因此，通常仅对其中一小部分进行检查，从而导致对栖息地连通性的描述较弱。为了解决这一计算难题，提出了一种多目标优化框架。首先详细介绍了可推广的多目标最小成本路径模型。然后介绍了一种用于确定多目标最小成本路径模型的受支持有效解决方案的非劣集估计（MONISE）算法。但是，众所周知，不受支持的有效解决方案（同等重要）也可以存在，但由于难以识别而通常会被忽略。从而，为了能够识别多目标模型的全套有效解决方案（支持和不支持），然后提出了一种多标准标记算法。已开发的框架可用于评估支持湿地系统中两栖动物运动的栖息地连通性的不同概念。结果强调了在表征连接性时需要权衡的范围，当认为多个目标有助于做出运动决策并且不受支持的有效解决方案（通常被忽略）的数量可能远远超过受支持的有效解决方案的数量时。已开发的框架可用于评估支持湿地系统中两栖动物运动的栖息地连通性的不同概念。结果强调了在表征连接性时需要权衡的范围，当认为多个目标有助于做出运动决策并且不受支持的有效解决方案（通常被忽略）的数量可能远远超过受支持的有效解决方案的数量时。已开发的框架可用于评估支持湿地系统中两栖动物运动的栖息地连通性的不同概念。结果强调了在表征连接性时需要权衡的范围，当认为多个目标有助于做出运动决策并且不受支持的有效解决方案（通常被忽略）的数量可能远远超过受支持的有效解决方案的数量时。