mysl / mm-chanshelton Goto Github PK

This is a study of An Electronic Market-Maker by Nicholas Tung Chan and Christian Shelton.

This is a Monte Carlo simulation of the basic model that computes the expected profit. The following is a reproduction of Figure 3 from the paper:

These are plots of the true price and market price for strategies 1, 2, and 3, respectively, when the noise factor alphaU = 0.4:

These are plots of the true price and market price for strategies 1, 2, and 3, respectively, when the noise factor alphaU = 1.0:

These are plots of the true price and market price for strategies 1, 2, and 3, respectively, when the noise factor alphaU = 1.6:

This is an implementation of the basic model that uses the SARSA learning method to choose the optimum strategy. The following is a reproduction of Figure 5 from the paper:

• Episode 25

• Episode 100

• Episode 200

• Episode 500

The following is a reproduction of Figure 6a from the paper:

The following is a reproduction of Figure 6b from the paper:

The following is a reproduction of Figure 6c from the paper:

This is an implementation of the extended model that uses the SARSA learning method to choose the optimum spread. The following is a reproduction of Figure 9 from the paper:

• Episode 25

• Episode 100

• Episode 200

• Episode 500

The following is a reproduction of Figure 10a from the paper:

The following is a reproduction of Figure 10b from the paper:

The following is a reproduction of Figure 10c from the paper:

The following is a reproduction of Figure 10d from the paper: