IMPLEMENTATION OF MATHEMATICAL OPTIMIZATION MODELS FOR TRANSPORT AND LOGISTICS MANAGEMENT: BARRIERS AND SUGGESTIONS

Mathematical optimization has been widely used in industries, business, and public administration. However, there has been criticism that mathematical optimization does not solve practical problems very well. This paper analyzes four barriers for implementing mathematical optimization models based on a real project we are involved in. We identify four barriers for implementing mathematical optimization models: (1) senior management may disagree on the decision variables, objective function, or constraints only at the last minute when the project is about to be completed; (2) the data provided by the client often contain errors; (3) a real project's objective function and constraints are often not well defined; and (4) optimization of an existing system is constrained by how the system has operated in the past. To overcome these barriers, we propose suggestions including communicating with senior management at early stages of a project, care-fully double-checking the data from the client, working closely with the client to identify the trade-offs in the project, and identifying what are allowed to be changed in an existing system before developing mathematical optimization models. Hopefully, this paper will be valuable for mathematical optimization specialists to solve real projects.