There wasnt suitable public domain software for Genetic Algorithms. So I wrote some the cost is that the program is for a specific application rather than a general shell. I have made the program available two ways:
The instructions for running the program via either method are available in the following page which is also linked via my Review page: www.lasalle.edu/~redmond/teach/456/GeneticAlgInstructions.htm . This is an optimization problem introduced in the Negnevitsky book; maintenance of power producing units is to be scheduled so as to maximize the lowest quarterly power surplus.
Task:
· Run the genetic algorithm program on the data: PowerUnitMaintHard.txt; try a variety of parameters if necessary until a successful solution is found. Note that a successful solution of the problem requires a solution to have a rating (strength) of at least zero (anything less and there is a period in which not enough power will be generated). I made this problem hard you may not get a successful solution right away.
a. Save all output (trace) files that you generate. Turn in soft (electronic) copies of the files.
b. Use Excel to graph the best and average solution strength by generation for the run that finds a solution. (you should be able to load the generation by generation results into Excel pretty easily). Turn in soft copy.
c. Copy a (highest rated) solution from the output panel. Turn in hard or soft copy.
d. Q1a: Explain the successful solution in a way a manager could understand.
e. Q1b: How many solutions were rated along the way? (NOT how many unique solutions; that would be VERY tedious to determine)
Parents:
solution: S476 { last2 first2 fourthQtr firstQtr fourthQtr secondQtr fourthQtr last3 last2 first3 secondQtr } (strength: -30.00)
solution: S491 { middle2 first2 firstQtr fourthQtr thirdQtr firstQtr thirdQtr first3 first2 last3 thirdQtr } (strength: -15.00)
Turn In:
Softcopies should be zipped together and submitted to Blackboard. Any hardcopies need to be turned in in person.
NOTE: