On The Allocation of Computer Educational Resources

Key ideas: Swarm theory, foraging theory, data acquisition, student objectives, student ability, group achievement, particle swarm optimization based fuzzy neural networks, decentralized decision making and self-organized systems, CAI.

This paper will address the question of whether swarm theory or foraging theory can improve the allocation of computer resources in education. This entails several component areas of education. The issue should be examined from the prospective of the student, the prospective of the class, the school, the school system and all within the prospective or capacity of the computers.

See Wikipedia for an analysis of the technology.

Background: When I first started writing educational lessons I wanted them to be as autonomous as possible. Obviously some lessons have prerequisites or foundation lessons, but within that constraint I wanted to be able to deliver lessons at random within specific subject areas.

I wanted a teacher to be able to assign 10% math, 10% spelling, 29% typing etc. The computer would randomly choose the lesson category from the assignment parameters and then randomly choose the individual lesson from that category of lessons where the student had completed all prerequisites.

The reason I wanted this flexibility was to test a theory about boredom. It is known that the mind stores many kinds of information in specialized areas or spread across several specialized areas. The question I wanted to answer was: “Is boredom related to the saturation or over use of such individual areas of the brain?” Would randomization of subject matter allow longer time on task by reducing saturation of specific areas of the brain?

The method of randomly selecting the next lesson could solve the problem for the test so long as a randomly chosen control group approached the same lessons sequentially.

A Better Method of Assignment: There may be a better way of assignment using swarm theory. Could each micro-lesson be like a bee or ant foraging for food? If the lesson recorded time for completion and accuracy of completion, resources could be allocated on the result plus a measure of importance for the lesson.

It would be poor policy to allocate all surplus resources to failed tasks ignoring areas of spectacular success. This would dumb down the individual with special talents.

Likewise it would be poor policy to allocate all surplus resources to areas of spectacular success because there are minimal skills that all individuals need to learn.

The algorithm should accomplish the task of proceeding rapidly where the student has a gift, but never letting the minimum skills be deficient.

If time were allocated say 25% to areas of success and 25% to areas of deficiency and 50% to all other subjects, the individual could proceed rapidly where the need or potential were greatest.

Rapid success could signal mastery of the subject matter which should trigger a diminution of delivery to a maintenance review. Persistent failure should trigger a regression to prerequisites.

In proceeding with a theme of successfulness or failure as triggers, mastery of a specific series of lessons could proceed into more difficult lessons of the same type or field with maintenance in the area of mastery. Similarly regression to prerequisites could dig deeper in the same field in face of persistent failure.

Jim Fuqua

---- End as of 19 August 2007 ---