Timely-Throughput Optimal Scheduling With Prediction

Abstract

Motivated by the increasing importance of providing delay-guaranteed services in general computing and communication systems, and the recent wide adoption of learning and prediction in network control, in this work, we consider a general stochastic single-server multi-user system and investigate the fundamental benefit of predictive scheduling in improving timely-throughput, being the rate of packets that are delivered to destinations before their deadlines. By adopting an error ratebased prediction model, we first derive a Markov decision process (MDP) solution to optimize the timely-throughput objective subject to an average resource consumption constraint. Based on a packet-level decomposition of the MDP, we explicitly characterize the optimal scheduling policy and rigorously quantify the timely-throughput improvement due to predictive-service, which scales as Θ(p C1 (a−amaxq), where a, amax, ρ ∈ (0, 1), C1 > 0, C2 ≥ 0 are constants, p is the truepositive rate in prediction, q is the false-negative rate, τ is the packet deadline and D is the prediction window size. We also conduct extensive simulations to validate our theoretical findings. Our results provide novel insights into how prediction and system parameters impact performance and provide useful guidelines for designing predictive low-latency control algorithms.

Minimizing Age of Information With Power Constraints: Multi-User Opportunistic Scheduling in Multi-State Time-Varying Channels

“…However the above work consider wireless fading to be an i.i.d process. When channel state evolution has Markov properties, scheduling to minimize delay performance and maximize throughput have been studied in [27], [28], [34]. Scheduling policy based on value iteration is proposed in [28] and a Whittle-like index policy to achieve delay-power trade-off is studied in [27].…”

“…When channel state evolution has Markov properties, scheduling to minimize delay performance and maximize throughput have been studied in [27], [28], [34]. Scheduling policy based on value iteration is proposed in [28] and a Whittle-like index policy to achieve delay-power trade-off is studied in [27]. In [34], the multi-user power and bandwidth constrained scheduling problem is solved by packet level decomposition, and an asymptotically optimum truncated scheduling policy is proposed.…”

“…• Inspired by [28], [34], [35], in Section IV-(A), we first relax the hard bandwidth constraint (5b) and adopt a sensor level decomposition by using Lagrange multiplier. After relaxation, multiple sensors can be scheduled simultaneously.…”

Scheduling to Minimize Age of Information in Multi-State Time-Varying Networks with Power Constraints

“…• Inspired by [19], [25], [27], we relax the hard bandwidth constraint (7b) into a time-average constraint and decouple the multi-user scheduling problem into single user CMDP. After relaxation, more than M users can be scheduled simultaneously.…”

“…with this equation the power constraint (7c) can be converted in the linear constraint (20f). The constraint Eqn, (20b)-(20d) can be obtained by substituting ξ x,q with y x,q and µ x , the relationship is obtained from (19). Notice that ξ x,q ≤ 1, the inequality constraint (20e) can be obtained.…”