案例库 · 研发与科研 · 技术决策 · 2012–2016
这条还没译成中文，下面是英文原文。
Syngenta used stochastic optimization to breed soybeans faster for $287M less
Syngenta modeled its soybean breeding pipeline with simulation and stochastic optimization, avoiding $287M+ in costs 2012–2016; won the 2015 Edelman Award.
Syngenta
那一手
Syngenta's soybean R&D faced a pipeline problem: with conventional breeding methods, crop productivity was not rising fast enough to meet demand, and every year of trials and trait introgression cost time and money. Choosing breeding strategies was largely judgment.
Syngenta codified its R&D process with advanced analytics: discrete-event and Monte Carlo simulation models represent the variety pipeline, and stochastic optimization picks breeding plans with high probability of success at minimum cost and time. Four tools — TI, BPL, YTD optimizer and DQC — support decisions from variety design to commercialization.
The new tools dramatically improved planning and decision-making, avoiding more than $287 million in soybean R&D costs from 2012–2016 while improving the odds of delivering a portfolio valued over $1.5 billion. INFORMS awarded Syngenta the 2015 Franz Edelman Award.
为什么管用
- Simulating the stochastic pipeline exposes which decisions actually drive cost and time.
- Optimizing the breeding plan beats optimizing each trial in isolation.
- Quantified probabilities make portfolio bets defensible and consistent.
可以搬走什么
When a long R&D pipeline is uncertain, simulate the whole process and optimize the plan, not just the experiments — savings come from better portfolios, not better trials.
后来呢
Syngenta applied the tools across soybean R&D, avoiding more than $287 million in costs from 2012–2016, and began a multi-year effort to roll similar analytics into all major crops. The program won the 2015 INFORMS Franz Edelman Award.
资料来源
- Syngenta Wins 2015 INFORMS Edelman Prize
- Advanced Analytics for Agricultural Product Development (Interfaces abstract)
- Syngenta wins 2015 INFORMS Franz Edelman Award
发现哪里写错了？告诉我们。