Abstract
A method of generating solutions for a broad range of problems associated with relational heuristics was developed. This system operates using an algorithm that generates balanced sets of agents with individualized goals and attributes. These agents are then assigned to fulfill the requirements of the system. The initial traits provided to the algorithm modify the partial query each individual agent makes on its initial breadth-first search for solutions pertaining the algorithm’s goals. The parameters subject to modification range from the depth the agent will search, to its willingness to select solutions that have already been explored by other agents. The results of each of these queries are then consolidated and examined by the algorithm using a scoring system that factors in agent attributes. All of these aspects result in a system that provides highly unique, yet structured solutions to a broad range of problems. The system is applicable in any scenario where a variety of distinct, heuristic answers are favored. Examples of these applications include recommendation systems, curriculum development for individuals or schools, or even medical diagnoses. The agent generation algorithm is highly configurable, allowing the system to adapt to a wide range of environments of varying complexity and size.