A robust and resilience machine learning for forecasting agri-food production

Antifragile, sustainable, and agile supply chain network design by considering resiliency, robustness, risk, and environmental requirements

This research suggests an Antifragile, Sustainable and Agile Supply Chain Network Design (ASASCND) as a new network design that integrates these concepts considering resiliency, robustness, risk, and environmental requirements. The cost function combines a novel method with robust stochastic optimization and Entropic Value at Risk (EVaR). This model combines expected value, maximum and EVaR of cost as an objective function. This research adds antifragility by the effect of learning on variable parameters, sustainability by considering the environmental and social issues, resiliency and agility by flexible capacity, and multi-resource and demand satisfaction constraints to the model. The case study is in the automotive industry. This model compares the main problem by considering antifragility without thinking about antifragility. The ASASCND cost is - 0.3% less than without considering antifragility. In addition, when the conservatism coefficient grows, the cost function increase. In addition, the antifragility coefficient and the confidence level affect positively, and the agility coefficient negatively affects the cost function. Expanding the model scale changes the cost function and time computation because the antifragility coefficient changes variable cost. Finally, managerial insights and practical implications are explained.

A robust, resilience multi-criteria decision-making with risk approach: a case study for renewable energy location

Regarding hard situations like war, the increasing cost of extraction and exploration of fossil fuels make governments move toward green and clear renewable energy (RE). As a result, we propose a novel multi-criteria decision-making (MCDM) method for RE location (REL) for the first time. This model suggests a Robust, Resilience MCDM with Risk approach (RRMCDMR) for REL. We propose a risk approach by adding a risk function in MCDM. A robust convex approach is used to tackle the uncertainty of the model for the real world. We compare the RRMCDMR problem in a wind farm location in Iran with different risk coefficient functions. As defined, Khaf, Nehbandan, and Esfarayan are in locations one to three in all modes. We changed the normalized risk function and suggested two other risk functions that can help risk-averse and risk-neutral decision-makers. We varied the robust convex coefficient and considered that by increasing the robust convex coefficient, the alternative score increased.