Recently, fine-scale structures have been constructed from initially large clusters of micron-size particles by inducing chaotic motion within cavities filled with polymer melts. Upon solidification, the structures can render the resulting composite material electrically conducting if conducting particles are used. The transport of particles and the formation of these structures by chaotic mixing in an eccentric cylindrical cavity is simulated computationally in order to assess the influence of key parameters. The general equation for particle motion in response to drag and colloidal forces was derived and dimensionless parameters were identified which characterize the process. Computational tracking of individual particles was performed under conditions where global chaotic mixing prevailed. Two levels of complexity were considered. In the first case, particles were assumed to move passively. In the second case, colloidal and drag forces were present. After mixing, networks comprising interconnected particles were identified as electrical pathways. Formation mechanisms are described at various stages of mixing and the networks are characterized statistically.