System and method for dynamic path- and state-dependent stochastic control allocation

• US 8,032,451 B2
• Filed: 02/14/2011
• Issued: 10/04/2011
• Est. Priority Date: 02/21/2007
• Status: Active Grant

• Current Assignee: David K.A. Mordecai
• Original Assignee: David K.A. Mordecai
• Inventors: Mordecai, David K. A.
• Primary Examiner(s): Merchant; Shahid
• Assistant Examiner(s): AMELUNXEN, BARBARA J
• Application Number:
  US13/026,927
• Publication Number:
  US 20110137792A1
• Time in Patent Office: 232 Days
• Field of Search: 705/35, 705/36, 705/36.R, 705/38, 703/23, 463/6
• US Class Current: 705/38
• CPC Class Codes: G06Q 20/10 specially adapted for electronic funds transfer [EFT] systems; specially adapted for home banking systems
• CPC Class Codes: G06Q 30/0238 at point-of-sale [POS]
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

The invention includes a system and process that employs contractual bargaining with agent-based computational methods for the dynamic allocation, optimization, and pricing of contingent rights and obligations between multiple counterparties with overlapping interests. The processes employ a dynamic and endogenous hierarchy or tiering of binding incentive compatible contingent strategies, which may include optimal liquidation policies for matched assets and liabilities based upon stochastic volume/price schedule related to statistically non-stationary supply/demand elasticities and order-flow, as well as variations in market microstructure. The invention includes a dynamic open system with distributed stochastic control of strategic interactions among dynamic optimizing agents across random states, wherein the actions of any one affects the joint costs and benefits for all the agents.

Claims

1. A method, the method comprising:

providing one or more processors operatively coupled to one or more processor readable media;

storing on the one or more processor readable media at least one database that includes;

electronic counterparty information representing a plurality of counterparties;

electronic control rights information representing respective rights to control inventory for at least two of the counterparties;

representing, by the one or more processors, each of the at least two counterparties by at least one respective agent;

observing, by at least one of the at least one respective agent for each of the at least two counterparties, at least one respective state;

mapping, by at least one of the at least one respective agent for each of the at least two counterparties, the at least one respective state to at least one probability of at least one conditional expectation formed in response to the at least one respective state;

mapping, by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one conditional expectation to at least one decision formed in response to the at least one conditional expectation;

mapping, by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one decision to at least one contingent payoff, and

outputting the at least one contingent payoff.

2. The method of claim 1, wherein at least one of the mapping the at least one respective state, the mapping the at least one conditional expectation and the mapping the at least one decision includes one or more of estimation, forecasting, inference, simulation and optimization operations.

3. The method of claim 1, further comprising determining, by the one or more processors, an optimal outcome across all of the at least one respective agent as a function of the contingent payoff.

4. The method of claim 1, wherein the outputting is made to a manifold.

5. The method of claim 1, further comprising:

receiving, by the one or more processors, a notification that at least two of the plurality of counterparties exchange respective paths;

simulating, by the one or more processors, propagated states and paths for the at least two counterparties; and

accounting, by the one or more processors, for evolving contractual rights and duties among the plurality of parties as a function of the propagated states and paths.

6. The method of claim 5, wherein the accounting comprises one or more of orthogonal Garch methodologies and Poisson shot-noise methodologies for estimating time-series properties of volatility and correlation.

7. The method of claim 5, wherein the accounting comprises Bayesian dynamic tree and kernel-based machine learning methodologies for updating state-dependent and path-dependent conditional probabilities and utilities.

8. The method of claim 5, wherein the simulating comprises employing, by the one or more processors, evolutionary computational methods that include one or more of evolution strategies, evolutionary algorithms, Bayesian networks, and state machines.

9. The method of claim 1, further comprising updating by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one conditional expectation as the at least one agent interacts in a state or with at least one other agent.

10. The method of claim 9, wherein the agents interact that include one or more of distributed searching, signaling, messaging, matching, and bargaining activities within a stochastic algorithmic game involving at least one agent.

11. A system, the system comprising:

one or more processors operatively coupled to one or more processor readable media;

at least one database stored on one or more of the one or more processor readable media, wherein the at least one database includes;

electronic counterparty information representing a plurality of counterparties;

electronic control rights information representing respective rights to control inventory for at least two of the counterparties;

wherein the one or more processors is programmed and configured to;

represent each of the at least two counterparties by at least one respective agent;

observe, by at least one of the at least one respective agent for each of the at least two counterparties, at least one respective state;

map, by at least one of the at least one respective agent for each of the at least two counterparties, the at least one respective state to at least one probability of at least one conditional expectation formed in response to the at least one respective state;

map, by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one conditional expectation to at least one decision formed in response to the at least one conditional expectation;

map, by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one decision to at least one contingent payoff, and

output the at least one contingent payoff.

12. The system of claim 11, wherein at least one of the mapping the at least one respective state, the mapping the at least one conditional expectation and the mapping the at least one decision includes one or more of estimation, forecasting, inference, simulation and optimization operations.

13. The system of claim 11, wherein the one or more processors is further programmed and configured to determine an optimal outcome across all of the at least one respective agent as a function of the contingent payoff.

14. The system of claim 11, wherein the outputting is made to a manifold.

15. The system of claim 11, wherein the one or more processors is further programmed and configured to:

receive a notification that at least two of the plurality of counterparties exchange respective paths;

simulate propagated states and paths for the at least two counterparties; and

account for evolving contractual rights and duties among the plurality of parties as a function of the propagated states and paths.

16. The system of claim 15, wherein the accounting comprises one or more of orthogonal Garch methodologies and Poisson shot-noise methodologies for estimating time-series properties of volatility and correlation.

17. The system of claim 15, wherein the accounting comprises Bayesian dynamic tree and kernel-based machine learning methodologies for updating state-dependent and path-dependent conditional probabilities and utilities.

18. The system of claim 15, wherein the one or more processors is further programmed and configured to employ evolutionary computational methods that include one or more of evolution strategies, evolutionary algorithms, Bayesian networks, and state machines.

19. The system of claim 11, wherein the one or more processors is further programmed and configured to update, by the at least one of the at least one respective agent for each of the at least two counterparties, the at least one conditional expectation as the at least one agent interacts in a state or with at least one other agent.

20. The system of claim 19, wherein the agents interact that include one or more of distributed searching, signaling, messaging, matching, and bargaining activities within a stochastic algorithmic game involving at least one agent.