1: \begin{abstract}

2: This paper discusses distributed optimization over a directed graph. We begin with some well known algorithms which achieve consensus among agents including Fast Row-stochastic-Optimization with uncoordinated STep-sizes (FROST) \cite{xin2019frost}, which possesses linear convergence to the optimum. However FROST works only over fixed topology of underlying network. Moreover the updates proposed therein require perfectly synchronized communication among nodes. Hence communication delays among nodes, which are inevitable due to processing delays and/ or channel impairments, preclude the possibility of implementing FROST in the real world. In this paper we utlize a co-operative control strategy which makes convergence to optimum impervious to communication delays.

3: \end{abstract}