In my PhD study, I addressed the challenges that limit higher data-densities and efficient energy utilization in application areas where low power consumption combined with dependable operation is required on NoC-based manycore system. Our research addresses the Dark Silicon (i.e., under-utilization of transistors on chip) problem directly through a set of energy-saving specialized cores which are monitored and managed by an efficient hierarchical agent-based approach. In my thesis we consider two different approaches to mitigate dark silicon viz. 1) a power management approach to consider an upper limit on total power consumption, dynamic behaviour of workloads, processing elements utilization, per-core power consumption, and load on network-on-chip and 2) mitigating dark silicon by utilizing dark areas for testing and enhancing power-aware online test scheduling method with explicit consideration of limited power budgets in many-core systems using runtime application mapping.