How can I use the bayesian network in RobotC ? Is possible ? Thanks Enzo

I suppose that you can write your BN program in RobotC mostly the same way you could do that in any other C implementation -- as long as your software fit in the memory. You can use integer as well as floating point variables etc.

There are performance limitations, however. Although RobotC seems to be much faster than standard Lego firmware (I have checked and confirm the claimed 20 x imrovement in execution speed) it is still a bytecode interpreter.

I also suggest to prepare a simulation of your problem and on a PC.