In addition to John George Hardy, the dispersion of filler within polymer matrix is contributed by type of polymer, type of filler (pure or modified one), drying temperature, and method of mixing.

thermodynamic determine the final structure of any system. time between initial state of system(that is changed with hydrodynamic and thermodynamic of system) and final state have priority. this time controls with hydrodynamic(kinetic) parameter of system like viscosity of media and thermodynamic parameters like particle surface tension.

We have to consider the complexity of the system. At least 6 parameters are effectively contributing to the final location of the nano-particles. 

Among kinetic parameters, viscosity ratio, mixing time, and mixing procedure ( i.e. premixing with one of the polymers and then adding the other one or mixing simultaneously) are of utmost importance. If viscosity ratio deviates largely from 1 it can play a strong role. To the best of my knowlage thermodynamics has highest priority unless you don't give enough time to the nanofiller to migrate to the preferentially interacting phase by precisely choosing the mixing conditions and capture the morphology in a temporary  state. You can consult this paper:

Fenouillot, Françoise, Philippe Cassagnau, and J-C. Majeste. "Uneven distribution of nanoparticles in immiscible fluids: Morphology development in polymer blends." Polymer 50.6 (2009): 1333-1350.

Here is another study done on MWCNT in polycarbonate and poly(styrene-acrylonitrile): DOI: 10.1002/marc.200800549

Apart from that I second Mr Hardy but in particular I want to stress the point Mrs Esnaashari and Mr Kushawaha make. "Complex" in this context can mean small deviations in the initial state (or procedure) can lead to large variations in the end. Thermodynamics certainly still applies but I would not expect a terribly tight correlation of results with TD for some nano particles. In my experience not only the large specific surface area but also their geometric shape can be a major influence. I have found carbon nanotube mixtures to be very sensitive to small changes, compared to e.g. carbon black.

Commenting on the influence of  shape and aspect ratio of  nanofiller , Goldel proposed the Slim-Fast-Mechanism (SFM) to explain why the transfer of “slim” nanofillers (CNT) occur significantly faster and more effective than that of “chubby” ones (carbon black) with the same wetting angle. Nanofillers with low aspect ratios are more likely to have a stable localization at the blend interface.

DOI: 10.1021/ma200793a

generally speaking, thermodynamic equilibrium localization can be predicted by Young's model, however, the kinetic parameters in some cases control the localization. The significance of the kinetic parameters depends on the system that you study (low or high interfacial tension, reactive or non-reactive, immiscible or partial miscible blend). 

You can find a detail discussion on the localization of solid particles in low and high interfacial tension blends in our following papers:

doi:10.1016/j.polymer.2015.08.046

doi:10.1016/j.polymer.2015.08.063

this article might help you

Studies on the selective localization of multi-walled carbon

nanotubes in blends of poly(vinylidene fluoride) and

polycaprolactone(DOI: 10.3144/expresspolymlett.2015.8)