Empirical Analysis on Market-Making Hedging Problem based on Quadratic Programming with Time-Decay and Sparse-Term

This paper considers a market-making hedging problem with real market demands. To minimize multi-day hedging error square, a procedure based on dynamic quadratic programming problem is designed. Considering that the confidence and predictive ability of time series attenuate with the length of interval, we add a time-decay factor to adjust sample influence. Besides, the smoothly clipped absolute deviations (SCAD) penalty term is applied to enhance the sparsity, and meanwhile remain the original properties of parameters. In addition, the economic meaning of solutions has been found for matching the industry distribution, which provides a reliable way to explain and verify the model practicability.