Dynamic oligopolistic competition on an electric power network with ramping costs and joint sales constraints

Reetabrata Mookherjee, Benjamin F. Hobbs, Terry L. Friesz, Matthew A. Rigdon

Research output: Contribution to journalArticlepeer-review

16 Scopus citations


Most previous Cournot-Nash models of competition among electricity generators have assumed a static perspective, resulting in finite dimensional variational and quasi-variational inequality formulations. However, these models' system costs and constraints fail to capture the dynamic nature of power networks. In this paper we propose a more general and complete model of Cournot-Nash competition on power networks that accounts for these features by including (i) explicit intra-day dynamics that describe the market's evolution from one Generalized Cournot-Nash Equilibrium to another for a 24 hour planning horizon, (ii) ramping constraints and costs for changing the power output of generators, and (iii) joint constraints that include variables from other generating companies within the profit maximization problems for individual generators. These joint constraints yield a generalized Nash equilibrium problem which can be represented as a dikerential quasi-variational inequality (DQVI); such generalized Nash equlibrium problems can have multiple solutions. The resulting formulation poses computational challenges that can cause traditional algorithms for DVIs to fail. A restricted formulation is proposed that can be solved by an implicit fixed point algorithm. A numerical example is provided.

Original languageEnglish (US)
Pages (from-to)425-452
Number of pages28
JournalJournal of Industrial and Management Optimization
Issue number3
StatePublished - 2008

All Science Journal Classification (ASJC) codes

  • Business and International Management
  • Strategy and Management
  • Control and Optimization
  • Applied Mathematics


Dive into the research topics of 'Dynamic oligopolistic competition on an electric power network with ramping costs and joint sales constraints'. Together they form a unique fingerprint.

Cite this