Demands for smaller lot sizes of mass-customized products increased the need for flexibility and adaptability in production lines. Semi-automatic manufacturing systems that involve human operators as well as technological equipment increase the flexibility of manufacturing systems. Such systems combine the benefits of human flexibility and new industrial and assistive technology. The key combinatorial problem to solve in the design of semi-automatic manufacturing lines is the assembly line balancing problem with the selection of equipment. An efficient and sustainable line design requires a cost-effective choice of equipment, and the presence of human increase the importance of ergonomics. In this work, we propose a Multi-objective Mixed-Integer Nonlinear Programming (MO-MINLP) for the design of semi-automated assembly lines. The objectives are the optimization of the design cost and the ergonomics level, modeled with the fatigue and recovery of workers. We propose to solve the problem with a bi-objective local search algorithm, based on the Iterative Local Search metaheuristic. We apply the algorithm on a case study to illustrate the originality of the problem and the solving algorithm.