Discrete-Event Control of Stochastic Networks: by Eitan Altman

By Eitan Altman

Beginning new instructions in examine in either discrete occasion dynamic platforms in addition to in stochastic regulate, this quantity makes a speciality of a large category of keep watch over and of optimization difficulties over sequences of integer numbers. this can be a counterpart of convex optimization within the environment of discrete optimization. the speculation constructed is utilized to the regulate of stochastic discrete-event dynamic platforms. a few purposes are admission, routing, carrier allocation and holiday keep watch over in queuing networks. natural and utilized mathematicians will take pleasure in examining the publication because it brings jointly many disciplines in arithmetic: combinatorics, stochastic techniques, stochastic keep watch over and optimization, discrete occasion dynamic structures, algebra.

Show description

Read or Download Discrete-Event Control of Stochastic Networks: Multimodularity and Regularity PDF

Similar linear programming books

Optimization Theory

"Optimization conception is turning into a a growing number of very important mathematical in addition to interdisciplinary quarter, in particular within the interaction among arithmetic and lots of different sciences like desktop technology, physics, engineering, operations learn, and so on. "This quantity supplies a accomplished creation into the idea of (deterministic) optimization on a sophisticated undergraduate and graduate point.

Trust-region methods

This is often the 1st complete reference on trust-region tools, a category of numerical algorithms for the answer of nonlinear convex optimization tools. Its unified remedy covers either unconstrained and restricted difficulties and experiences a wide a part of the really good literature at the topic.

Convex analysis

On hand for the 1st time in paperback, R. Tyrrell Rockafellar's vintage research offers readers with a coherent department of nonlinear mathematical research that's particularly suited for the research of optimization difficulties. Rockafellar's concept differs from classical research in that differentiability assumptions are changed through convexity assumptions.

Hybrid Dynamical Systems : Modeling, Stability, and Robustness

Hybrid dynamical structures convey non-stop and immediate alterations, having positive aspects of continuous-time and discrete-time dynamical platforms. full of a wealth of examples to demonstrate suggestions, this e-book offers a whole thought of sturdy asymptotic balance for hybrid dynamical structures that's appropriate to the layout of hybrid keep an eye on algorithms--algorithms that characteristic common sense, timers, or mixtures of electronic and analog elements.

Additional resources for Discrete-Event Control of Stochastic Networks: Multimodularity and Regularity

Example text

It is sufficient to show that any set Θ ⊂ AX sequentially normally compact at y¯ (with respect to the whole space Y ) is also SNC at y¯ with respect to the smaller Banach space AX . 22 for the surjective operator A: X → AX . 21 ensuring that codim AX < ∞ due to aff Θ ⊂ AX . , there is a closed subspace Z ⊂ Y with AX Z = Y . Now denote by Nε (·; Θ| AX ) the set of ε-normals to Θ with respect to AX and take Θ arbitrary sequences yk → y¯, εk ↓ 0, and yk∗ ∈ Nεk (yk ; Θ| AX ) converging to ∗ zero in the weak topology of (AX )∗ .

We have yˆ∗ ∈ Y ∗ . Since ∇ f (¯ ∗ remains to prove that yˆ ∈ Nc2 ε (¯ y ; Θ) with some constant c2 > 0. To furnish this, we use again the metric regularity property for the mapping f and its strict derivative. 16) for f with some µ > 0, we find x y ∈ f −1 (y) such that x y − x¯ ≤ µ y − y¯ . 16) for the operator ∇ f (¯ x ), we get xˆy ∈ ∇ f (¯ x )−1 (y − y¯) with 22 1 Generalized Differentiation in Banach Spaces x y − x¯ − xˆy = o( x y − x¯ ) . Now putting all the above together, one has lim sup Θ y →¯ y xˆ∗ , xˆy xˆ∗ , xˆy yˆ∗ , y − y¯ = lim sup ≤ lim sup max 0, −1 y − y¯ y − y¯ µ x y − x¯ Θ Θ y →¯ y y →¯ y = lim sup max 0, Θ y →¯ y ≤ µ lim sup max 0, ε + x f −1 (Θ) → ¯ x xˆ∗ , x y − x¯ µ−1 x y − x¯ x ∗ , x − x¯ x − x¯ ≤ 2µε .

The principal difference between tangential and normal approximations is that the former constructions provide local approximations of sets in primal spaces, while the latter ones are defined in dual spaces carrying “dual” information for the study of local behavior. Being applied to epigraphs of extended-real-valued functions and graphs of set-valued mappings, tangential approximations generate corresponding directional derivatives/subderivatives of functions and graphical derivatives of mappings, while normal approximations relate to subdifferentials and coderivatives, respectively; see below.

Download PDF sample

Rated 4.97 of 5 – based on 6 votes