The final project for this course will be comprised of three parts: one mandatory problem and two more problems chosen from the four options given. You are required to submit a short write-up for each problem, making three sections of one single final report. Make sure your final report contains all the basic elements of a technical report, i.e., a proper (short) introduction to the subject, explanation of the method, the raw results, discussion of the results, and any final remarks or observations. 1. (Mandatory) Genetic algorithm for the Traveling Salesman Problem (TSP). Use the collection of GA subroutines you wrote for the homework assignments to write a complete GA that solves the TSP problem. Use appropriately customized versions of the crossover, mutation and other genetic operators you wrote, and perhaps write new ones, to solve the 10-, 100-, 1000- and 10,000-city versions of the problem. Use the random (r. y) coordinates on [0, 1] given with a random (twister) seed of 1. Submit the results in the form of a table showing the shortest path and solution time for each number of cities. Also show a plot of the optimum solution for the first two cases only. Use popsize=round(3(n)), where n is the number of cities, and genmax-round (30(n)). Set the various probabilities as you see fit. 2. (Possible) Constrained, multi-objective evolutionary algorithm (MOEA). Use the Matlab toolbox to solve the constrained, multiobjective optimization problems given below. Submit the driver routine you wrote and any necessary functions/subroutine you used. Be sure to display the Pareto surface at the end of the evolution. 1 (x) = 25 (11 2) (12 2) (3 1) (11 4) (rs 1) Minimize= s.t. = 2(x) = 91 (2) 1+2 220 92 (r) 93 (a) 6-12 20 2-20 |g4 (2) 2-1 +32 20 gs (x)=4-(3-3) - 4 0 96 (x) = (-3)+1-420 where 0.2, 16) 10, 1 {3, 5) 5, and 0 6. 3. (Possible) Particle Swarm Optimization for a constrained problems (PSO). Modify the PSO code you have wrote for your homework to use the penalty method to solve the con- strained, single-objective Mishra's Bird function optimization problem. Use the constraint < 5 instead of