Machine-to-Machine (M2M) communication is the emerging technology for automation applications without human intervention. The 3GPP organization defines Machine Type Communications (MTC) to enhance support of M2M devices. Due to a potentially massive number of coexisting MTC devices, achieving efficient device identifier (ID) management and signaling is challenging. In this paper, a scheduled device grouping scheme is proposed to minimize the number of required device IDs and signaling overhead. The grouped devices share one ID while sequentially attaching to the network to execute application specific tasks. Applications with various goals are fulfilled considering expiration time and task dependency. Beyond formulating an integer programming problem, the issue is mapped into a variation of bin packing problem. The solution reaches 30% to 50% ID reduction by adopting a classic fitting algorithm.