Configuration and operation of building transportation systems, e.g., elevators and stairs for offices, hotels, and apartments, are important, and have profound societal impact such as in improving efficiency, reducing costs, and saving lives. Establishing methodologies that are effective and coherent across configuration and operation phases while covering both normal and emergency modes, however, is difficult. In this paper, coherent configuration and operation of building transportation systems for both normal and emergency modes are studied through a synergistic integration of optimization, formal semantics, and constraint satisfaction. Based on a formal semantics, a statistical configuration method using a coarse-grain model and an optimization-based operation method using a fine-grain model are developed. These methods are integrated by using constraint programming to efficiently select high quality configurations with performance coherent across to the operation phase for both normal and emergency modes.