Discussion

The proposed DRX market clearing mechanism follows the notion of a competitive market as was defined by general equilibrium theory in microeconomics. In particular, it seeks to explain the competitive behavior of buying and selling agents in a DRX market, and then proving that equilibrium prices for all DR quantities can be achieved via a tˆatonnement process under such agent behavior.

The theoretical work developed in this paper should be regarded as an extension of those presented in microeconomics. In fact, general equilibrium theory [75] focuses inten-

Figure 4.14: Newton tˆatonnement with various ranges of estimation error

Figure 4.15:Newton tˆatonnement via finite-difference online estimation ofJt

sively on the treatment of markets for private goods. To the best of our knowledge, very little attention has been paid to developing a similar (complete) theory for competitive trading of public goods (such as national defense, fresh air, common lands, rivers, etc). This lack of attention is explained by the fact that provision of most of these goods are usu- ally under a government control through the use of taxation. Consequently, a competitive negotiation scheme for rights to use these goods does not appear to be practical.

However, DR as a form of public goods is different. The provision of DR isnot subject to government intervention. In fact, DR can be produced by any customer whenever they are paid to do so. Also, DR can be supplied to any buying agents (Recos, Transcos, Discos)

Summary

whenever they need it and are willing to pay. In this sense, developing a competitive market for the trading of DR as public goods is both feasible and necessary.

Based on this review of general equilibrium theory, we suggest that our DRX proposal for a public good market clearing mechanism, using either classical or Newton price tˆa- tonnement methods, makes a sound contribution to the field of microeconomics. We also believe that the theoretical framework developed in this paper for competitive trading of DR can also be applied to trading other public goods, as long as they are not under government control. In fact, the framework is generic in the sense that it involves multiple buyers, multiple sellers, and multiple products or quantities (i.e., an “exchange economy”). Note also that the framework did not make any assumptions beyond those common in microeconomics, such as the “preference convexity” made for proving the existence and the uniqueness of market equilibrium.

4.7

Summary

This chapter presented the design and evaluation of an agent-based market clearing scheme for the eXchange of Demand Response. The proposed scheme uses Walrasian auctions, where participating agents update their quantity bids in response to prices adjusted by the market operator. This auction is repeated iteratively until the market equilibrium is obtained at a point where the market outcome is Pareto optimal. Both the existence and the uniqueness of this equilibrium are proven under the condition that agent preferences are strictly convex.

The price adjustment is performed using either classical or Newton tˆatonnements. Both methods have advantages and limitations. Although classical tˆatonnement is easy to im- plement, it may not converge to an equilibrium solution if the value of price adjustment factor K is not suitably chosen (i.e., within its feasible range). On the other hand, the Newton method offers robust convergence, although in return it requires greater computa- tional effort in estimating the Jacobian matrix online. This estimation is performed using the finite-difference principle.

Cost–Benefit Analysis and

Treatment of Externalities

5.1

Overview

This chapter proposes a comprehensive framework for assessing short–term financial costs and benefits derived from scheduling DR using various market-clearing schemes, including the conventional partial schemes and the DRX introduced in previous chapters. Through- out the framework development, we analyse DR cost and benefit for each participant in a DR(X) market, as well as for Gencos and Recos in the wholesale electricity market. Based on these local analyses, a global evaluation is performed to determine whether the optimised DR can give a positive social surplus. If so, the DR will be dispatched during the time period (i.e., an hour) under consideration. The proposed cost–benefit assessment framework is illustrated on a small power system, and its usefulness reported through the demonstration of externalities across all involved parties. Case studies on the Roy Billinton Test System (RBTS) are given to examine the scalability of the proposed framework.

One of the main motivations for developing this framework is to establish a rigorous

evaluation of DRX as the key innovation proposed this thesis. Here the evaluation relies on only a few justifiable and, in fact, justified assumptions which demonstrate the validity of the market outcome including DR costs and benefits derived from a DRX clearing scheme. In addition to this, the framework can be used to evaluate certain classes of conventional DR schemes such as the Transco–based, the Reco–based, and the Disco– based. This substantiates the applicability of the proposed assessment framework.

The chapter is structured as follows. Section 5.2 introduces methodologies commonly used for an economic assessment. Based in these standard methods, the overall structure of the proposed framework for evaluating DR costs and benefits are described in Section 5.3, with the details given in subsequent sections, 5.4 and 5.5. The developed framework is illustrated via a small-scale study in Section 5.6, and then its scalability is examined in Section 5.7 using the RBTS. Conclusing remarks are finally given in Section 5.8.

Common economic assessment methodologies