Consumer model

• US 6,741,973 B1
• Filed: 10/14/1997
• Issued: 05/25/2004
• Est. Priority Date: 04/04/1997
• Status: Expired due to Term

• Current Assignee: NCR Corporation
• Original Assignee: NCR Corporation
• Inventors: Dove, Lee G., Johnson, Graham I.
• Primary Examiner(s): Davis, George B.
• Application Number:
  US08/950,230
• Time in Patent Office: 2,415 Days
• Field of Search: 706/13, 706/47
• US Class Current: 706/13
• CPC Class Codes: G06N 3/126 Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

A model of consumer behavior in a transaction environment such as customers moving around a bank branch, is generated from an artificial life algorithm to create a number of agents. In each agent, a genetically encoded drive, equivalent for example to hunger, is defined so as to correspond to a transaction need such as the need for cash. Interaction rules, such as navigation rules, are set for interaction between the agents and a first representation of an environment, and the program is run and the agents observed, then compared with real human behavior. The best matched agents are selected and the program run again, the steps being repeated until a required level of comparison with real behavior is reached. The model can then be used with different transaction environments to study customer behavior and to select the best branch layout or the like.

Claims

1. A method of generating a model of customer behavior in a transaction environment, comprising by the steps of:

(a) selecting a software development tool incorporating at least one artificial life algorithm and capable of constructing a plurality of agents each having at least one drive;

(b) defining at least one drive for each agent which is matched to a transaction-related need; and

(c) genetically encoding the defined drives.

2. A method according to claim 1, further comprising the steps of:

(d) generating a representation of a transaction environment; and

(e) generating a plurality of agents each having a drive corresponding to a transaction-related need and each capable of interacting with the environment in accordance with predetermined rules.

3. A method according to claim 2, further comprising the steps of:

(f) causing the agents to interact with the representation of a transaction environment;

(g) recording the interactions;

(h) comparing the interactions of each agent with the representation of the environment against real life interaction of humans with the transaction environment; and

(i) selecting a number of agents which provide the best comparison.

4. A method according to claim 3, steps (f), (h), and (i) are repeated until a required number of agents conform to a predetermined level of comparison.

5. A method of predicting and exploring potential patterns of customer behavior in a transaction environment, the method comprising the steps of:

(a) generating a customer behavior model which includes a plurality of agents each having a drive corresponding to a transaction-related need and each capable of interacting with the environment in accordance with predetermined rules;

(b) generating a representation of a different transaction environment;

(c) causing selected agents to interact with the different representation; and

(d) observing the interactions of step (c).

6. A method of modeling human behavior, comprising the following steps:

a) creating a computer model of a financial institution;

b) creating a group of agent-models which undergo interactions with

i) each other and

ii) the financial institution;

c) monitoring the interactions of the agent-models;

d) monitoring actual interactions of actual people with

i) a financial institution; and

ii) each other; and

e) using the actual interactions to select agent-models from the group.

7. Method according to claim 6, and further comprising the step of

f) running simulations using the selected agent-models, and not using non-selected agent-models.

8. A method of modeling customers in a financial environment, comprising the following steps:

a) creating a number of model-agents,

i) each equipped with at least one drive, and

ii) some of the drives comprising a need for cash;

b) creating a model of a financial environment;

c) creating a set of interaction rules, which govern interactions

i) among model-agents and

ii) between model-agents and the environment;

d) running the model, wherein the model-agents perform interactions

i) among themselves and

ii) with the environment;

e) comparing the interactions of paragraph (d) with actual interactions

i) among actual persons and

ii) between actual persons and an actual financial environment; and

f) selecting the model-agents whose model interactions most closely resemble the actual interactions.

9. Method according to claim 8, and further comprising the step of running a simulation in which the model-agents selected in paragraph (f) are used in a simulation involving a different financial environment.

10. A method of modeling customer behavior in a financial environment, comprising the following steps:

a) running a first simulation, wherein a group of simulated people perform simulated behaviors, wherein they interact with each other and with a first financial environment;

b) selecting simulated people from the group; and

c) running a second simulation, wherein the selected simulated people of paragraph (b) perform simulated behaviors in a second financial environment.

11. Method according to claim 10, wherein the selection of paragraph (b) is based at least partly on comparison of behavior of the simulated behaviors with actual behavior of actual people.