Optimization Approaches to Multi-robot Planning and Scheduling

The study and use of multi-robot teams has become more prevalent within academia and industry as the capability and autonomy of these systems continues to improve (Arai, Pagello, and Parker 2002). With high levels of progress already made concerning the control of individual robots, the acknowledged advantages of multi-robot systems (MRS) has resulted in considerable research attention in the last couple of decades in efforts to build more efficient systems for their coordination (Gerkey and Matarić 2004). These advantages include resolving more complex tasks, increasing the speed at which tasks can be completed, and enhancing the level of system reliability and redundancy present within singlerobot solutions (Yan, Jouandeau, and Cherif 2013). This thesis, though in early stages of development, concerns the field of study within MRS known as multi-robot task allocation (MRTA) (Gerkey and Matarić 2004). Specifically, it proposes to investigate the integration of techniques from the optimization literature, namely mixedinteger and constraint programming, within architectures for MRTA. This research area aims to solve multi-robot coordination problems pertaining to task distribution to robot resources and the temporal scheduling of tasks on such resources. These problems have a wide variety of real-world applications including planetary exploration (Mataric and Sukhatme 2001), airport and harbor transhipment (Alami et al. 1998), and emergency response (Østergård, Matarić, and Sukhatme 2001).