Georgia Institute of Technology College of Sciences

I will talk about the problem of computing the number of integer partitions into parts lying in some integer sequence. We prove that for certain classes of infinite sequences the number of associated partitions of an input N can be computed in time polynomial in its bit size, log N. Special cases include binary partitions (i.e. partitions into powers of two) that have a key connection with Cayley compositions and polytopes. Some questions related to algebraic differential equations for partition sequences will also be discussed. (This is joint work with Igor Pak.)

In the latent voter model, which models the spread of a technology through a social network, individuals who have just changed their choice have a latent period, which is exponential with rate λ during which they will not buy a new device. We study site and edge versions of this model on random graphs generated by a configuration model in which the degrees d(x) have 3 ≤ d(x) ≤ M. We show that if the number of vertices n → ∞ and log n << λn << n then the latent voter model has a quasi-stationary state in which each opinion has probability ≈ 1/2 and persists in this state for a time that is ≥ nm for any m <∞. Thus, even a very small latent period drastically changes the behavior of the voter model.

Optimization problems arising in decentralized multi-agent systems have gained significant attention in the context of cyber-physical, communication, power, and robotic networks combined with privacy preservation, distributed data mining and processing issues. The distributed nature of the problems is inherent due to partial knowledge of the problem data (i.e., a portion of the cost function or a subset of the constraints is known to different entities in the system), which necessitates costly communications among neighboring agents. In this talk, we present a new class of decentralized first-order methods for nonsmooth and stochastic optimization problems which can significantly reduce the number of inter-node communications. Our major contribution is the development of decentralized communication sliding methods, which can skip inter-node communications while agents solve the primal subproblems iteratively through linearizations of their local objective functions.This is a joint work with Guanghui (George) Lan and Yi Zhou.