Guayule and guar are two desert-dwelling crops that can provide raw materials year-round for bioproducts such as rubber, resin, guar gum, and guar meal. Both crops are low-water-use, drought-tolerant, as well as heat-resistant, and these features enable their great potential for the agricultural economy in the Southwestern U.S. However, there exist challenges when considering the design of their supply chains in not only the economic benefits but also the environmental and social impacts, such as the process facility location and transportation problems. Furthermore, the optimal supply chains are heavily dependent on the amount of crop production, which can be measured by the adoption rate, i.e., the percentage of current crops in the field that is switched to either guayule or guar. In this paper, stochastic scenarios are utilized to capture the uncertainties of the adoption rates of each field. Afterward, a stochastic optimization is deployed to identify optimal decisions for facility locations, transportations from fields to facilities, and finally to customers, with a multi-objective approach to quantify the economic benefits (minimizing the costs of supply chains), environmental impacts (minimizing CO${}_2$ equivalent greenhouse gas emissions), and social impacts (maximizing the local accrued jobs). Based on the Geographic Information System for capturing field information and relevant factors, and deciding facility locations, the model is formulated as a complex large-scale mixed-integer linear optimization problem. For an efficient solution, the Benders Decomposition algorithm is implemented. The proposed approaches are evaluated based on the cases of two areas: Maricopa and Pinal counties in Arizona for the guayule supply chain, and Dona Ana County in New Mexico for the guar supply chain.

Guayule和瓜尔豆是两种荒漠化的农作物，可以全年为橡胶，树脂，瓜尔豆胶和瓜尔豆粉等生物产品提供原材料。两种农作物都耗水少，耐旱且耐热，这些特性使其在美国西南部的农业经济中具有巨大的潜力。然而，在考虑设计供应链时存在挑战。不仅是经济利益，还包括环境和社会影响，例如过程设施的位置和运输问题。此外，最佳供应链在很大程度上取决于农作物的产量，这可以通过采用率来衡量，即通过转换为番石榴或瓜尔豆的田间当前农作物的百分比来衡量。在本文中，随机情景被用来捕获每个领域采用率的不确定性。然后，采用多目标方法量化经济效益（使供应链成本最小化），对环境的影响（最小化），以随机优化的方式确定设施位置，从田间到设施以及最终到客户的运输的最佳决策。一氧化碳${}_2$等效的温室气体排放量）和社会影响（最大限度地增加当地应计工作）。基于用于捕获现场信息和相关因素的地理信息系统，并确定设施位置，该模型被公式化为一个复杂的大型混合整数线性优化问题。对于有效的解决方案，实施了Benders分解算法。根据两个区域的情况对提议的方法进行了评估：亚利桑那州的Maricopa和Pinal县用于瓜尤尔牛供应链，新瓜纳州的Dona Ana县用于瓜尔豆供应链。