This thesis is comprised of three major projects of my research. In the first project, I proposed a nanoparticle model and combined it with the Theoretically Informed Coarse Grained (TICG) model for pure polymer systems and the grand canonical slip springs model developed in our group to build a new model for entangled nanocomposites. With Molecule Dynamics(MD) simulation, I studied the mechanic properties of the nanocomposites, for example the influence of nanoparticles size and volume fraction on entanglements, the diffusion of polymers and nanoparticles, and the influence of nanoparticles size and volume fraction on viscosity et al.. We found that the addition of small-size nanoparticles reduces the viscosity of the nanocomposites, which is in contrary to what Einstein predicted a century ago. However, when particle increases its size to micrometers the Einstein predictions is recovered. From our simulation, we believe that small-size nanoparticles can more effectively decrease the entanglements of nanocomposites than larger particles. The free volume effect introduced by small-size nanoparticles also helps decrease the viscosity of the whole system. In the second project, I combined the Ohta-Kawasaki (OK) model  and the Covariance Matrix Adaptation Evolutionary Strategy(CMA-ES) to optimize the block copolymer blends self-assembly in the hole-shrink process. The aim is to predict the optimal composition and the optimal surface energy to direct the block copolymer blends self-assembly process in the confined hole. After optimization in the OK model, we calibrated the optimal results by the more reliable TICG model and got the same morphology. By comparing different optimization process, we found that the homopolymers which are comprised of the same monomers as either block of the block copolymer can form a perfect perforated hole and might have better performance than the pure block copolymer. While homopolymers which are comprised of a third-party monomers have negative effects on forming the perforated hole regardless of the composition and the surface energy. In the third project, We applied the OK model and the CMA-ES algorithm to optimize the combined outcome of particular substrate-copolymer combinations, and we arrived at an efficient method for design of substrates leading to nontrivial desirable outcomes.