sf logistics' "fengzhi" big model makes supply chain decision-making easier
2024-09-21
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
image description: gao lei
our reporter ding yazhi
in the complex and highly specialized field of logistics supply chain, the application of big models is accelerating. at the services trade fair, sf express brought its self-developed breakthrough results, the "fengzhi" logistics decision-making big model, which was just released in august, and the industry's first full-link logistics carbon footprint management system, the "fenghe sustainable development platform".
the "fengzhi" model can be divided into three modules. gao lei, chief engineer of sf intelligent algorithms, introduced that the first one is the supply chain intelligent agent, which mainly helps customers to optimize the supply chain analysis and diagnosis. when out-of-stock affects sales, sf technology's supply chain intelligent agent can respond quickly and tell users what caused the out-of-stock and the corresponding improvement strategies to cope with market changes.
at the same time, based on the multimodal large model, sf technology has further built a multi-scale and multi-channel prediction model to help customers better predict future needs, thereby improving planning accuracy and resource efficiency.
the third module is the logistics decision-making model. "we know that the process of generating text by models such as chatgpt is actually a process of generating a word-by-word text sequence, and many problems in logistics can actually be regarded as a sequence generation problem. for example, if i want to deliver to 100 stores today, i need to plan a route, first to a, then to b, to complete the delivery at the lowest cost. therefore, we can use some of the same underlying technologies as the language model to directly generate logistics routes, thereby breaking through the limitations of the original large model language mode and directly solving the core logistics decision-making problems of users." gao lei said. in practical applications, the large model significantly accelerates the computational efficiency of enterprise packaging optimization and transportation route planning, helping enterprises achieve the optimal solution for overall fulfillment timeliness and transportation costs.
in addition, sf express has built a standard carbon management system and developed its own digital carbon management platform - "fenghe sustainable development platform", which is used to integrate the carbon emissions and carbon reduction data of the entire group, covering packaging, transportation, transit, delivery and other links.
gao lei said: "many of sf express's partners, target customers and peers also have a strong sense of social responsibility. we hope that after the sustainable platform is productized, it will enable more companies to optimize carbon management, reduce carbon emissions and achieve emission reduction targets."▲# hundreds of reviews#
