Azure Data Science Virtual Machines (DSVM) vs. IBM CPLEX Optimization Studio

Azure DSVM is useful in [a] Machine Learning environment where GPU-based processing is [required]. [The] most relevant [users] for the Azure DSVM is in ML/AI for model training and processing [high-end] CPU tasks with GPU compatibility. Azure DSVM is built for [a] startup to low medium IT environments where the ML/AI-based projects are [carried] out.

It is well suited for solving large-sized, mixed-integer, and integer programming problems. Now, the new version supports for Multi-Objective optimization along with some new algorithms such as Benders Decomposition. It is less appropriate for quadratic programming problems where the objective function is the product of multiple variables. However, it's very easy to code any problem.

• Leveraging data.
• Computer vision.
• Data science.

• Integer programming.
• Mixed integer programming.
• Quadratic programming.
• Multi-objective optimization.

• Azure DSVM pricing must be reduced so that an AI-based start-up can use the Azure DSVM.
• Azure must create an environment to use Azure DSVM offline as well.
• Lack of frameworks

• Data handling from different sources like Note Pad, etc.
• Large size of MILP problems.
• Various parameters to set.

It's nice to use and with good optimization.

Honestly, to say, I never contacted CPLEX but used its forum to know/clarify any issues I faced.

It's within the Azure environment and it's easy to manage.

Compared with MATLAB, CPLEX is a more user-friendly and simpler structure for writing models. This one also has a good return on investment.

• Azure DSVM is little costly with long term support for ML based environments.
• Azure DSVM is very good for short tasking and costs us [a] little low than the on-prem server.
• [Scaling] option is very convenient.

• Better for price.
• Many model parameters/features.
• No visualization.