Updates

Why OPL is important for modelling

Robert Ashford Optimization Direct
and Alex Fleischer IBM

Modelling tools help users capture their business environment or organizational situation as a mathematical model. Such models provide insight and control that cannot be gained elsewhere. The most powerful models are optimization models and many businesses such as petrochemicals, supply chain, airlines and logistics cannot operate competitively without them.

Increasing computer power and data availability have made optimization models an effective tool in almost every commercial and organizational activity. Building, running and maintaining such models is a significant task. IBM’s OPL and CPLEX and CP software make it much easier. OPL enables the user to build models using the natural mathematical formulation to specify the model structure in terms of variables and inequations, read data from files and spreadsheets, and use it to populate specific instances of the situation of interest. CPLEX and IBM’s CP optimizer are the most powerful solvers around and work hand-in-glove with OPL to optimize the model efficiently and supply it with solutions which OPL can analyse.

The process is made even more effective by IBM’s ILOG CPLEX Optimization Studio – a development environment for OPL, CPLEX and the CP optimizer, which makes these tools very easy to use. An interactive editor provides ‘tool tips’ advice, parameters and options are accessible through control panels and there is full, searchable interactive help. Debugging is made easy both by the editor which identifies errors before execution and by the Optimization Studio’s ability to link execution issues to specific model objects and data.

OPL itself is a non-procedural language for specifying optimization models and is intended and designed for the analyst rather than the programmer. Models expressed in it are self documenting and intelligible to the non-specialist. It is ideal for prototyping and developing models to the production stage. A late 2017 survey of IBM’s optimization users revealed that more than 40% rely on OPL.

Whilst it is easy to get started modelling and using optimization with the Optimization Studio, it is possible to build and manage sophisticated applications using OPL’s javascript language: models can be solved iteratively or in parts and multiple scenarios can be run and managed. And once development is sufficiently advanced, OPL and the optimizers may be run from command line scripts or Python.

The framework provided by OPL makes it easy for many users to understand, maintain and develop the same model, so the models and applications in which they can be embedded readily survive personnel changes and achieve organizational longevity.

Computationally demanding optimization tasks can be run on cloud based servers and the Optimization Studio provides a seamless way of doing this. Going a step further it is even possible to use IBM provided APIs to call OPL and CPLEX or the CP optimizer entirely in the cloud.

Contact Optimization Direct to find out more.