Agent awareness modeling for agent-based modeling systems

1. A method for modeling agent awareness in an agent based model for simulating human behavior, the method comprising the steps of:

defining a plurality of agents;

assigning a plurality of attributes to each agent of the plurality of agents, wherein the plurality of attributes includes at least one attribute value selected from an attribute group consisting of age, race, income, TV watching habits, Internet usage habits, radio listening habits, price sensitivity, historical purchase occasions, initial awareness, and quality sensitivity;

for each agent, generating a uniformly-distributed, blended variable value for values of at least two of the attributes in the attribute group that are uniformly-distributed by blending;

for each agent, assigning the generated blended variable value as an additional attribute value;

tracking a ratio of indicators for each agent;

varying the ratio of indicators for each agent upon the occurrence of a triggering event for that agent, based on at least one of the values of the plurality of attributes in the attribute group and the blended variable for that agent;

using the ratio as a factor to model the agent's awareness as a continuous, non-discrete value

running a series of test simulations in which at least one variable is adjusted by an amount while the remaining variables remain constant;

determining a change in mean absolute percentage error (“MAPE”) of each of the test simulations;

for each of the test simulations, automatically optimizing the values by at least adjusting values of variables that cause the MAPE to change by an amount that exceeds an error for the test simulation to optimize the simulation of the human behavior;

utilizing the values of the variables to control simulation of the human behavior by the agents;

simulating human behavior by the agents including agents making decisions and reacting to input stimuli according to individual characteristics and constraints of each agent;

generating simulated human behavior output data from the simulation agents; and

applying the simulated human behavior output data to determine and implement a real-world plan to exploit expected human behavior based on the simulated human behavior.

2. The method of claim 1 wherein the triggering event is an advertising event, or a purchasing event, or a distribution event.

3. The method of claim 1 wherein the triggering event is the passage of time.

4. The method of claim 1 wherein the ratio of indicators for each agent relates to at least two types of indicators, and wherein the method tracks the number of each of said at least two types of indicators associated with each agent, and wherein each triggering event for an agent causes a predetermined number of at least one type of indicators to be removed from, or added to, the number of that type of indicators associated with the associated agent.

5. The method of claim 1 wherein the varying step is repeated for each of the plurality of agents until beta distribution of ratio of indicators is reached across the agents.

6. The method of claim 1 wherein the using step includes utilizing the ratio to adjust a purchase probability of the agent.

7. The method of claim 1 wherein the maintaining and varying steps are used to model using step includes modeling agent awareness utilizing the Polya Um method.

8. The method of claim 1 further including the steps of defining a plurality of agents, assigning a plurality of attributes to each agent, wherein one of the attributes is initial awareness, the method further including running a simulation in which each agent is introduced to a stimuli and determining the reaction of each agent to the stimuli.

9. The method of claim 8 wherein the running a simulation step is an agent-based simulation which predicts consumer responses to advertising or promotional efforts.

10. The method of claim 8 wherein the awareness attribute models the agent's awareness of a particular item that is available for purchase.

11. The method of claim 1 wherein the method is implemented on a processor which carries out the tracking, varying and using steps.

12. The method of claim 1 wherein the agent's awareness takes the form of an awareness frequency ratio representing the percentage of stimuli or other cues which the agent encounters which cause the agent to think about a given brand.

13. A method for running an agent-based simulation for simulating human behavior comprising:

defining a plurality of agents;

assigning a plurality of attributes to each agent, wherein one of the attributes is a non binary continuous, non-discrete awareness attribute that models the agent's awareness of a particular item that is available for purchase, and wherein the plurality of attributes further includes at least one attribute value selected from an attribute group consisting of age, race, income, TV watching habits, Internet usage habits, radio listening habits, price sensitivity, historical purchase occasions, initial awareness, and quality sensitivity;

for each agent, generating a uniformly-distributed, blended variable value for values of at least two of the attributes in the attribute group that are uniformly-distributed by blending;

for each agent, assigning the generated blended variable value as an additional attribute value;

running a series of test simulations in which at least one variable is adjusted by an amount while the remaining variables remain constant;

determining a change in mean absolute percentage error (“MAPE”) of each of the test simulations;

for each of the test simulations, automatically optimizing the values by at least adjusting values of variables that cause the MAPE to change by an amount that exceeds an error for the test simulation to optimize the simulation of the human behavior; and

utilizing the values of the variables to control simulation of the human behavior by the agents;

in an actual simulation of the human behavior, introducing each agent to a stimulus such that each agent reacts to the stimulus at least partially based upon values of the agent's assigned attributes; and

determining or recording data representing the reaction of each agent to the stimulus

applying the determined or recorded data to determine and implement a real-world plan to exploit expected human behavior based on the simulated human behavior.

14. The method of claim 13 wherein the awareness attribute is determined by tracking a ratio of indicators for each agent and varying the ratio of indicators for an agent upon the occurrence of a triggering event for that agent.

15. The method of claim 14 wherein the triggering event is an advertising event, or a purchasing event, or a distribution event, or the passage of time.

16. The method of claim 13 wherein the awareness attribute relates to awareness of a particular brand, and is an unbounded variable that is related to the number of mental associations the agent has with the brand.

17. The method of claim 13 wherein the awareness attribute is adjusted to reflect a reduced awareness solely due to the passage of time.

18. The method of claim 1 wherein the plurality of attributes includes an initial awareness attribute determined by sampling a probability function.