To run a MD simulation in any software (gromacs, namd, lammps, dl-poly etc.) you need the partial charges if previously not included in the topology file.
You can use quantum software packages like Gaussian or Gamess.
To make your force field compatible with oplsaa, you need to look at the mother paper of that force field and get information about how they have done (i.e. methods) the quantum parts.
Or, you can run with MP2/6-31G++(d,p) with NBO calculation and extract the electrostatic/geometric information. You can see the log file to get charges. Don't go for Mullikan charges. I don't know how big your molecule is, based on that and the computational resources available, you can move on to higher level quantum calculations (coupled cluster/casscf ). You may not follow the exact calculations by those of opls developers, It won't kill your simulation I guess.
You may not change the .rtp file. Make a copy of the forcefield folder to the working directory and edit there. pdb2gmx is not a must. you can start without it bit the .itp files has to be included in a customised topol.top file with #include command.