Optimization of truck tyres selection

Abstract: This thesis, which consists of an introduction and five appended papers, concerns the optimal selection of tyres for a variety of vehicle configurations as well as operating environments. The selection problem stems from a project cooperation between Chalmers University of Technology and Volvo Group Trucks Technology. We analyze the selection problem from a mathematical optimization point of view. The overall purpose is to reduce the tractive energy required to run the vehicle. We develop a computationally efficient vehicle dynamics model of the vehicle, the tyres, and the operating environment. The tyres are represented by a surrogate model of the rolling resistance coefficient, which measures the energy losses caused by the tyres. The properties of the surrogate model called for a methodology for connecting expert knowledge about a general simulation-based function with its radial basis function interpolation. An algorithm for the solution of a large set of instances of a simulation-based optimization problem with continuous variables has been developed and tested on a set of problem instances. This algorithm enables an efficient computation of approximately optimal tyre designs (represented by continuous variables) for each vehicle configuration and operating environment specification. A splitting algorithm for simulation-based optimization problems with categorical variables has been developed and evaluated on a set of test problems. This algorithm outperforms all algorithms applicable to this class of optimization problems, and finds an approximately optimal tyres configuration. Since each execution of this algorithm requires many computationally expensive evaluations of the simulation-based objective function, it cannot be used to solve the full tyres selection problem. The two latter algorithms are then combined to enable the efficient solution of many instances of a simulation-based optimization problem with categorical variables. The resulting algorithm is applied to a couple of instances of the tyres selection problem. Our experiments show that the optimization methodology developed enables a computationally efficient solution of the truck tyres selection problem, in the combinatorial domain of possible vehicle configurations and operating environment specifications. Putting our methodology into practice will involve many challenges besides the problems studied in this thesis; however we have shown that our methodology can be utilized in the sales tool at Volvo.