As the number of autonomous decision-making entities in the electricity grid increases, it is necessary to develop (1) new decision-making capabilities embedded within the grid's control and management, and (2) new grid architecture models ensuring that both individual and system objectives are met. This work develops (1) new decision-making mechanisms enabling residential energy users and electricity providers to interact through the use of dynamic price signals, and (2) policy recommendations to facilitate the emergence of shared architecture models describing the future state of the electricity grid. In the first part, two optimization models that capture the emerging flexible consumption, storage, and generation capabilities of residential end-users are formulated. An economic dispatch model that explicitly accounts for end-users' internal dynamics is proposed. A non-iterative pricing algorithm using convex and inverse linear programming is developed to induce autonomous residential end-users to behave cooperatively and minimize the provider's generation costs. In the second part, several factors that make the development of grid architecture models necessary from a public policy standpoint are identified and discussed. The grid architecture problem is rigorously framed as both a market failure legitimizing government intervention, and a meta-problem requiring the development of non-conventional methods of solution. A policy approach drawing on the theoretical concepts of broker, boundary object and boundary organization is proposed.