Titlebook: Stochastic Network Optimization with Application to Communication and Queueing Systems; Michael J. Neely Book 2010 Springer Nature Switzer

贝雷帽

Optimization of Renewal Systems, the frame. A policy is a contingency plan for making a sequence of decisions, where new random events might take place after each decision in the sequence. This model allows a larger class of problems to be treated, including . described in more detail in Section 7.6.2.

FRONT

Introduction,is on communication and queueing systems. Example applications include wireless mesh networks with opportunistic scheduling, cognitive radio networks, ad-hoc mobile networks, internets with peer-to-peer communication, and sensor networks with joint compression and transmission. The techniques are al

偏离

Introduction to Queues,s assumed to be a non-negative real valued random variable. Future states are driven by stochastic arrival and server processes . and . according to the following dynamic equation:.We call . the . on slot ., as it can represent an amount of work that needs to be done. The stochastic processes . and.

BILK

Dynamic Scheduling Example,pts for a simple 2-user wireless downlink example, similar to the example given in Section 2.3.2 of the previous chapter. First, the problem is formulated in terms of known arrival rates and channel state probabilities. However, rather than using a randomized scheduling algorithm that bases decision

Mitigate

Optimizing Functions of Time Averages,on-convex functions of time averages. To begin, consider a discrete time queueing system .(.) = (.1(.)‚…,.(.)) with the standard update equation:.Let . be attribute vectors. As before, the arrival, service, and attribute variables are determined by general functions . and ..Consider now the followin

macrophage

Approximate Scheduling,rference. Previous chapters showed the key step is maximizing the expectation of a weighted sum of link transmission rates, or coming within an additive constant . of the maximum. Specifically, consider a (possibly multi-hop) network with . links, and let . be the transmission rate offered over link

执

Optimization of Renewal Systems,very slot . a single random event .(.) is observed, a single action α(.) is taken, and the combination of α(.) and ω(.) generates a vector of attributes (i.e., either penalties or rewards) for that slot. Here, we change the slot structure to a . structure. The frame durations are variable and can de

虚度

Conclusions,oped to solve these problems. Our focus was on communication and queueing networks, including networks with wireless links and mobile devices. The theory can be used for networks with a variety of goals and functionalities, such as networks with:

就职

synovium

978-3-031-79994-5Springer Nature Switzerland AG 2010