CPLEX 12.10 on WML
Read how straightforward it is to redeploy optimization models on WML and benefit from the latest CPLEX 12.10 improvements.
IBM CPLEX Optimization Studio 12.10 is available and offers several improvements. Among them, performance improvements in the CPLEX and CP Optimizer engines. With Decision Optimization for Watson Machine Learning, this is very easy to get these benefits as you just need to redeploy the model with the 12.10 runtime.
Simply change runtime and model type
As seen in the DO for WML REST API documentation, the software specifications (including versions of engines used to solve your problems) are stated at the time you create the deployed model on WML. The given payload looks like:
curl -X POST \ https://us-south.ml.cloud.ibm.com/v4/models \ -H "Authorization: bearer IAM-Token-here" \ -H "ML-Instance-ID:ML-InstanceId-here" \ -H "Content-Type:application/json" \ -H "cache-control:no-cache" \ -d "{\"name\":\"Diet\", \"description\":\"Diet\", \"type\":\"do-docplex_12.9\",\"runtime\": {\"href\":\"/v4/runtimes/do_12.9\"}}"This payload can be modified as follow to use the latest version of CPLEX:
curl -X POST \ https://us-south.ml.cloud.ibm.com/v4/models \ -H "Authorization: bearer IAM-Token-here" \ -H "ML-Instance-ID:ML-InstanceId-here" \ -H "Content-Type:application/json" \ -H "cache-control:no-cache" \ -d "{\"name\":\"Diet\", \"description\":\"Diet\", \"type\":\"do-docplex_12.10\",\"runtime\": {\"href\":\"/v4/runtimes/do_12.10\"}}"The version on the model type indicates the version of the modeling API used, as docplex (and other modeling API) might evolve over time.
The version of the runtime indicates the version of the engines to be used.
You are done!
You can now use your deployments from your application, simply using this new deployment ID. Note that you can even have two alternative deployments at the same time and use both and validate the outcomes are the same.
