






The objectives of the series of assignments are multifold:
 To use the TSP as a test bed for various ML approaches (e.g., genetic algorithms, simulated annealing, Kohonen Self Organizing Maps etc).
 To briefly study various real world applications of the TSP (e.g., logistics; genome sequencing; manufacturing: IC testing, PCB drilling; aiming telescopes; and xray crystallography)
 To tie in problems from machine learning to general Computer Science (e.g., revisiting the concepts of NP completeness).
 To provide a historical perspective of a problem with a rich history, significant practical applications, and and numerous technological approaches.)
 To use the algorithm(s) developed for solving the TSP to solve a different optimization problem (predicting the DOW).













Conceptually the requirements of this assignment are simple. The assignment as given in my class required familiarity with Python programming. Additionally students need to be familiar with basic prediction methods (kNN, neural nets etc) and basic optimization techniques (genetic algorithms etc).















This exercise can be given in a couple of different variations (1) in a short form as described in the SIGCSE nifty assignment or (2) in a longer form as described in the above pdf file TSP.pdf. You will need the free Adobe Acrobat Reader to view this file.















The course syllabus is available here.






