html5-img
1 / 36

Berthing Problem

Berthing Problem. Chen Fang Yew Nicholas 1 , Gani Zhi Hao Terry 1 , Vo Thanh Minh Tue 1 1 NUS High School of Mathematics and Science, Singapore. Introduction. The given problem is very complex and arises in daily management of the Port of Singapore .

toril
Download Presentation

Berthing Problem

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Berthing Problem Chen Fang Yew Nicholas1, Gani Zhi Hao Terry1, Vo Thanh Minh Tue1 1NUS High School of Mathematics and Science, Singapore

  2. Introduction • The given problem is very complex and arises in daily management of the Port of Singapore. • The port success depends on a robust and efficient berthing plan. • The program must compute on demand.

  3. Introduction • We propose a greedy first fit algorithm to solve the berthing problem. • We also introduce graph theory as a possible approach to solve the problem.

  4. The problem • Ships arrive at various times and have different lengths and berthing times. • Our task is to devise a berthing plan that minimizes the waiting time for all ships.

  5. Objective • Minimize the waiting time function • Where is the time the ship is berthed, Is the arrival time, and is the waiting constant

  6. Wait time scenario 1

  7. Wait time scenario 1

  8. Wait time case study • When =1, LHS=RHS • Either order is the same.

  9. Wait time scenario 2

  10. Wait time scenario 2

  11. Wait time case study • Second option incurs more waiting time • Minimizing total height equates to minimizing total j, waiting time. • First fit algorithm.

  12. First fit algorithm • Greedy algorithm. • Choose the lowest time possible. • Choose the lowest space possible.

  13. Sample data

  14. Advantages • Extremely fast to compute. • Give a reasonable good solution. • Modify the algorithm to meet realistic constraints.

  15. Modified realistic algorithm • “port stay times are often delayed beyond the estimated values” • Takes into account delay time • Models after realistic conditions • Allows for inter-clearance distance between ships

  16. Discretization • Transform continuous data into discrete values. • Simplify the data input. • Scale the solution range.

  17. Sample data

  18. A novel approach • Apply graph theory. • Rigorous mathematical ground.

  19. Berthing problem • Variables • ai: Arrival time of ith ship. • hi: Processing time. • Li: Length of the ship. • Waiting penalty for ith ship: (yi – ai )α

  20. Berthing problem • Solution: { (t1, y1), (t2, y2), …, (tn, yn) } • ti: Time in which ith ship starts to dock. • yi : Lower y-coordinate of ith ship. • Objective function.

  21. Solution space

  22. Solution space • (ti, yi) • Every point in a solution space is a feasible solution. • Overlapping of solution spaces yields solution domain.

  23. Solution space • Overlapping of rectangle <=> incompatible solution pairs. • Compatible solution pairs. • (ti, yi)  (tj, yj)

  24. Solution space

  25. Graph theory • Each solution (ti, yi, i) can be assigned as a vertex. • Compatible solutions are joined by an edge. Vertex for ith ship: (ti, yi) E = (ti – ai)α + (tj – aj)α Vertex for jth ship: (tj, yj)

  26. Graph theory • Consider six ships A to F with solution pairs SA to SF • The vertices A to F and their edges form a complete graph • A cycle is created as all the vertices join to each other once. AF A F AB EF E B BC DE CD D C

  27. Graph theory • By adding the weight of each edge of the cycle together, the overall delay time can be calculated.

  28. Graph theory • A swap in position of vertices within the cycle does not change the overall delay time. AF AF A F F A AB BF EF AE = E E B B BC DE BC DE CD CD

  29. Graph theory • Consider another vertex not within the cycle • If weight of new edge < weight of old edge, replace old vertex with new vertex AF A F FG AB EF G E B BC DE DG CD

  30. Graph theory • For example, • If wFG + wDG < wEF + wDE, then replace E with G AF A F FG AB EF G E B BC DE DG CD

  31. The algorithm • Start with the earliest ship. • Depth-first-search for a possible complete sub graph. • Check all vertices in the graph if improvement possible. • If no improvement possible, then the program terminates.

  32. Advantages • Greedy algorithm can give a reasonable good solution. • By transforming from a geometrical problem to a graph problem, we can handle constraints more easily.

  33. Challenges • Large number of vertices and edges. • Complex relationship.

  34. Conclusion • Justified using first fit algorithm • Devised an algorithm to solve the problem • Improved on it to take into account realistic conditions • Suggested a novel approach using graph theory to solve this problem

  35. References • Dai, J., Lin, W., Moorthy, R., & Teo, C.-P. (2003). Berth Allocation Planning Optimization in Container Terminal. 1-33. • Duin, C. W., & Sluis, E. v. (n.d.). On the complexity of Adjacent Resource Scheduling. • Lim, A. (n.d.). An Effective Ship Berthing Algorithm.

  36. The end • Thank you for your attention.

More Related