Recently, an alternative approach to dynamic nuclear polarization (DNP) in the liquid state was introduced using optical illumination instead of microwave pumping. By exciting a suitable dye to the triplet state which undergoes a diffusive encounter with a persistent radical forming a quartet-doublet pair in the encounter complex, dynamic electron polarization (DEP) is generated via the radical-triplet pair mechanism. Subsequent cross-relaxation generates nuclear polarization without the need for microwave saturation of the electronic transitions. Here, we present a theoretical justification for the initial experimental results by means of numerical simulations. These allow investigation of the effects of various experimental parameters, such as radical and dye concentrations, sample geometry, and laser power, on the DNP enhancement factors, providing targets for experimental optimization. It is predicted that reducing the sample volume will result in larger enhancements by permitting a higher concentration of triplets in a sample of increased optical density. We also explore the effects of the pulsed laser rather than continuous-wave illumination, rationalizing the failure to observe the optical DNP effect under illumination conditions common to DEP experiments. Examining the influence of the illumination duty cycle, the conditions necessary to permit the use of pulsed illumination without compromising signal enhancement are determined, which may reduce undesirable laser heating effects. This first simulation of the optical DNP method therefore underpins the further development of the technology.