The successful implementation of the electricity market model has challenged the conventional way of operating the power system. In the electricity market model, power system is restructured to promote private companies to participate in a market structure where companies can sign a binding contract with large customers or can participate into pool market structure. Generation companies (GENCOs) and customers submit their bids in blocks in a pool market structure. GENCOs can achieve profit through strategic bidding due to the competitive nature of market structure. For this objective to realize, historical data of bidding of other participants should be modeled. This chapter addresses the application of dynamic programming, game theory and various AI based tools to form strategic bidding in the real time electricity market. To extend the analysis, a comparison of methods of designing bidding strategies has been presented. Based on this comparison, a critical review has been carried out to investigate the leading methods of strategic bidding.
TopIntroduction
Conventional power systems, consisting of vertical integrated utility suffer from rigidity, organizational complexity, management difficulties and risk of failure. The regulations which were followed earlier in these systems are not much relevant at the present time. Advanced technologies, research, integration of renewable sources and enhancement in security of power system and revenues demanded deregulation of power system. Deregulation or restructuring of power system required leniency in government policies and encouragement of private firms to participate in the operation of power system (David & Wen, 2000). This led to inception of market structure where public and/or private firms participate to sell produced power in blocks. These evolutionary changes are depicted in the timeline diagram in Fig. 1. Customer participation is of great significance here to improve energy efficiency and reliability. Large customers can play a vital role in the market structure as they can bid to purchase electricity. In a pool market structure, the market operator records the bid submitted by market participants and determines Market Clearing Price (MCP) (David, 1993). MCP is an equilibrium price determined from the supply and demand curve and it is the maximum price at which energy can be sold or purchased. Prices remain close or moving towards MCP in a competitive market structure. But due to the oligopolistic nature of electricity markets, large companies can exercise market power and influence MCP.
Figure 1. Evolution of deregulated power system
GENCOs can also enter into long-term bilateral transactions with large customers. There are many benefits yet there are many challenges in horizontal structure also such as electricity markets are not perfectly competitive. Only a few power generation companies can participate, which indicates the oligopolistic nature of market structure. In this oligopolistic pool market structure, GENCOs submit their bid to Market Operator who dispatches the feasible trading solution (Ansari & Rahimi-Kian, 2015)
GENCOs can increase their profits by predicting rival GENCO’s behavior. Historical data is needed to model this complex problem. The profit, which is to be maximized, can be calculated by:
(1)The price and quantity offered are denoted by Cji and Pji respectively where i is bid in blocks for jth generator. Product of these two is noted as revenue. 𝜑j denotes profit and generator cost is given by GCj.
To gain profit, GENCOs have to be aware of their rival’s behavior which can be predicted using normal Probability Density Function. The distribution of prices of bids can be modeled as:
(2)
is the price which is a random variable here with mean value 
and standard deviation 
. The function represents distribution of prices of rival GENCOs.
Based on this knowledge, GENCOs can strategically plan their bid using different approaches addressed by authors in the past such as Dynamic programming, Stochastic Optimization and Game Theory approach. Deep learning-based methods have gained wide recognition recently due to their robust and reliable nature. An approach based on multi agent system has also been reviewed to show the efficacy of these methods. A review of supply side and demand side bidding strategies has also been presented.