1: \begin{abstract}

2:     This paper considers the economic dispatch problem for a network

3:     of power generating units communicating over a strongly connected,

4:     weight-balanced digraph.  The collective aim is to meet a power

5:     demand while respecting individual generator constraints and

6:     minimizing the total generation cost.  We design a distributed

7:     coordination algorithm consisting of two interconnected dynamical

8:     systems. One block uses dynamic average consensus to estimate the

9:     evolving mismatch in load satisfaction given the generation levels

10:     of the units. The other block adjusts the generation levels based

11:     on the optimization objective and the estimate of the load

12:     mismatch.  Our convergence analysis shows that the resulting

13:     strategy provably converges to the solution of the dispatch

14:     problem starting from any initial power allocation, and therefore

15:     does not require any specific procedure for initialization.  We

16:     also characterize the algorithm robustness properties against the

17:     addition and deletion of units (capturing scenarios with

18:     intermittent power generation) and its ability to track

19:     time-varying loads.  Our technical approach employs a novel

20:     refinement of the LaSalle Invariance Principle for differential

21:     inclusions, that we also establish and is of independent interest.

22:     % Our technical approach uses notions and tools from algebraic

23:     % graph theory, nonsmooth analysis, set-valued dynamical systems,

24:     % and dynamic average consensus.

25:     Several simulations illustrate our results.

26:   \end{abstract}

27: