How to solve an optimization problem >4 attributes

I am working on an optimization problem that exists of diverse formulas and attributes. The type of problem is comparable with the field of operations research. The challenge is to find a set of attributes with an optimal outcome. I tried to solve this type of problem with the parameter optimization loop. For an easy problem with for instance two formulas and iteration over 4 attributes this already takes much time. Adding attributes and formulas will increase the time exponentially. I have to use Brute force while the Hillclimbing option doesn’t give the right solution. I solved the performance problem by generating a list with all possible combinations of values and calculate the formula’s in parallel by using the table row to variable loop. But this leads to very long tables that exceed more than 100 million rows for 8 attributes. Two questions on this:

  • Does Knime offers other kinds of support for the field of Operations research (comparable to the solver function in Excel)?
  • I currently generate the table with all possible values of the attributes with the parameter optimization loop. But for 100 million rows I need this number of iterations which is very time consuming. Is a smarter way of generating this table possible?

I'm not aware of any builtin functionality for Operations research, however, there are several Java libraries that seem to have this functionality. One option might be to use KNIME Analytics Platform to prepare all your data and then use the Java Snippet node to import the library and calculate the values that you need.

Libraries I found are: 


Thanks for your answer. This requires additional expertise in Java. Python also offers OR libraries, R seems less comprehensive in this field.