Achieving reasonable solving performance with a model can sometimes be a challenging task, even with help from the world's fastest optimization engine. The ILOG OPL-CPLEX Development System makes model performance tuning easier.

Managing data sets, parameter settings and models
The first step in performance tuning is to create a set of scenarios, combining particular alternate model formulations, data instances and engine parameters. OPL's Run Configurations feature is specifically designed to support the systematic isolation of performance problems.

Run Configuration allows you to organize models and data sets into executable scenarios. Enlarge image (opens a new window).

When an OPL project contains multiple models with multiple engine settings and multiple data files, a Run Configuration references the specific combinations to be executed. Easily configurable through drag-and-drop, a benchmark suite is easy to manage, right from within the OPL integrated development environment (IDE).

The Profiler in OPL-CPLEX helps you identify the precise bottlenecks in your models, including showing how memory and time is spent, both during model generation and model execution.

Example from the Profiler, detailing time and memory consumption for a MILP model. Enlarge image (opens a new window).