Multiprocessor Scheduling Part 6

Tham khảo tài liệu 'multiprocessor scheduling part 6', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 140 Multiprocessor Scheduling Theory and Applications V Y is not worse than the current best solution . RECORD plus a controlled DEVIATION. The algorithm of record-to-record travel to minimization may be generalized as follows RRT for minimization generate an initial solution choose an initial DEVIATION 0 set RECORD the current best solution value Repeat generate a neighborhood solution that is a perturbation of the current solution Y pErtuRb X IF V Y RECORD DEVIATION THEN accept the move . X Y IF V Y RECORD THEN set RECORD V Y IF no improvement on the solution quality after a number of iterations THeN lower DEVIATION IF the stop criterion is reached THeN stop GOTO Repeat Tabu lists Since neighborhood solutions not leading to improvement are accepted in RRT it is possible to return to previously visited solutions and cause cycling solutions. Hence tabu lists from the tabu search Glover 1989 are applied to overcome this problem. The tabu lists store attributes that identify certain moves are forbidden in the later search. By using tabu lists the solutions previously searched may be avoided and new regions of the search space may be explored. Theoretically the tabu lists need to store all previously visited solutions. However this would require too much memory and computational efforts. An practical way is to store only the moves occurring in the last s iterations in which s is known as the tabu size. By using an appropriate tabu size the likelihood of cycling solutions may be avoided. An aspiration criterion is used to free a tabu solution if it is of sufficient quality and possibly would not cause cycling solutions. Hence a solution is not forbidden if its attributes are not tabu or it passes the aspiration criterion test. 4. Heuristic procedure to minimize Cmax and computational results The proposed heuristic to minimize Cmax Heu_Cmax and computational results are presented in this section. The development of Heu_Cmax is based on observing secondary

Không thể tạo bản xem trước, hãy bấm tải xuống
Đã phát hiện trình chặn quảng cáo AdBlock
Trang web này phụ thuộc vào doanh thu từ số lần hiển thị quảng cáo để tồn tại. Vui lòng tắt trình chặn quảng cáo của bạn hoặc tạm dừng tính năng chặn quảng cáo cho trang web này.