关键词: 行为运筹学, 泊位分配, 岸桥分配, 二阶锥规划, 外逼近

Abstract: In terms of the basic ideas of behavioral operation research, a service-oriented Mixed-Integer Nonlinear Programming (MINLP) model by using a deep integration method was formulated to deal with the berth allocation and quay crane assignment problem. To overcome the problems of computational intractability and optimality absence, the model was transformed into a Mixed-Integer Second Order Cone Programming (MISOCP) model, and the optimization software CPLEX was used to solve the model. Since the branch and cut algorithm was time-consuming and ran out of memory for some instances, MINLP model was solved by outer approximation algorithm. Based on the characteristics of the problem, the model was decomposed into a mixed-integer linear programming master problem and a nonlinear programming sub-problem. The optimal solution to the sub-problem was obtained by analytical method, and the outer approximation of nonlinear constraints was derived by using the first-order Taylor series expansion. Finally, solution performance of the branch and cut algorithm and the outer approximation algorithm was compared by numerical experiments. Moreover, convergence of the outer approximation algorithm was verified. The sensitivity analysis of key parameters in the model was also conducted.

Key words: behavioral operation research, berth allocation, quay crane assignment, second order cone programming, outer approximation, scheduling