Minimizing GC work by analysis of live objects

As complex embedded real-time systems are widely used, automatic memory management, or garbage collection, is a technique that significantly reduces the complex problem of correct memory management. In recent years, people often use time-based real-time GC to meet the deadline of real-time tasks. Some algorithms are proposed to reduce memory requirement based on time-based approaches. However, they mainly focus on scheduling strategy and GC algorithms such as copying algorithm and mark-sweep algorithm et al. This paper treats GC as a dynamical process and gives the analysis of amount of live objects. We propose that GC should process memory space of each task one by one based on the variance of amount of live objects. The proposed approach can minimize GC work and further reduce memory requirement.