Resource Optimisation at Maritime Oil Terminals

The problem we consider in this thesis is the discrete dynamic berth allocation problem with integrated pipeline assignment planning. In this problem process- ing speed of oil batches depend on the chosen set of pipeline segments to the goal tank as pipeline segments may have dierent maximum pumping speeds. Cleaning time can also be taken into account if this is needed when sequentially processing dierent oil products through the same pipeline segment. We successfully come up with a linear mixed integer programming formu- lation to describe the considered problem. We show that for cases of realistic sizes CPLEX may not be able to solve the problem to optimality. Selecting the strong branching variable selection technique results in lower upper bounds if the problem cannot be solved to optimality. Strong branching mostly results in optimality faster too. The heuristics that we implement are the benchmark rst-come, rst-served (FCFS) heuristic and squeaky wheel optimisation to further improve on the initial FCFS solution. A comparison of the heuristics and the exact method with seven test cases shows that CPLEX with strong branching results in better upper bounds but sometimes also needs more than an hour to solve to optimality. The heuristics on the other hand run in negligible time.