The encyclopedia · R&D & Science · Operational decision · 1998–2004
Dow AgroSciences schedules new-product R&D with simulation-based optimisation.
Dow AgroSciences paired simulation with heuristic search to schedule new-product development, saving several million dollars.
Dow AgroSciences
the move
Dow AgroSciences had to schedule a whole portfolio of new products through R&D, where activities share scarce resources and order constraints.
The team examined mathematical programming for the job but found the formulations intractable at practical size once resource constraints were included.
Instead of giving up, they built a simulation-based framework to generate feasible schedules, then used heuristic optimisation such as tabu search and genetic algorithms over the precedence relations among activities.
Implemented as a Six Sigma project, the approach saved several million dollars from 1998 to 2004 and reduced the time and cost of moving its product portfolio to market.
why it works
- Exact math programming could not handle the resource constraints at real size
- Simulation made infeasible schedules detectable quickly
- Heuristics operating on precedence relations kept the search structured
- A scheduling win shortened time-to-market, compounding the value
what transfers
When an exact model is too big, simulate feasibility and let a heuristic explore the precedence structure — you can still reach strong schedules without solving the exact integer program.
what came after
Implemented via a Six Sigma project, the simulation-based optimisation saved Dow AgroSciences several million dollars from 1998 to 2004 and reduced the time and expense of bringing its product portfolio to market. The Interfaces (2004) paper became a reference for R&D pipeline scheduling.
references
- Dow AgroSciences uses simulation-based optimization to schedule the new-product development process
- Dow AgroSciences Uses Simulation-Based Optimization to Schedule the New-Product Development Process
spotted an error? The archive wants to know.