# The achievable region method in the optimal control of queueing systems; formulations, bounds and policies

@article{Bertsimas1995TheAR, title={The achievable region method in the optimal control of queueing systems; formulations, bounds and policies}, author={Dimitris Bertsimas}, journal={Queueing Systems}, year={1995}, volume={21}, pages={337-389} }

We survey a new approach that the author and his co-workers have developed to formulate stochastic control problems (predominantly queueing systems) asmathematical programming problems. The central idea is to characterize the region of achievable performance in a stochastic control problem, i.e., find linear or nonlinear constraints on the performance vectors that all policies satisfy. We present linear and nonlinear relaxations of the performance space for the following problems: Indexable…

## 47 Citations

### Optimization of Multiclass Queueing Networks with Changeover Times Via the Achievable Region Approach: Part I, the Single-Station Case

- Computer Science
- 1999

This approach seeks to obtain performance bounds and scheduling policies from the solution of a mathematical program over a relaxation of the system's performance region to address the performance optimization problem in a single-station multiclass queueing network with changeover times.

### Optimization of Multiclass Queueing Networks with Changeover Times Via the Achievable Region Method: Part Ii, the Multi-Station Case

- Computer Science
- 1998

We address the problem of scheduling a multi-station multiclass queueing network (MQNET) with server changeover times to minimize steady-state mean job holding costs. We present new lower bounds on…

### Strongly asymptotically optimal design and control of production and service systems

- Computer Science
- 2000

A formal definition of strong asymptotic optimality in the context of design and control of queueing systems that can be modeled as single or multiple stage queueing networks is provided.

### The achievable region approach to the optimal control of stochastic systems

- Computer Science
- 1999

The achievable region approach seeks solutions to stochastic optimization problems by characterizing the space of all possible performances (the achievable region) of the system of interest and…

### A linear programming approach to stability , optimisation and performance analysis for Markovian multiclass queueing networks ★

- Business
- 1999

Our object of study is a multiclass queueing network (MQNET) which consists of a collection of (connected) single-server stations. Exogenous arrivals into the system form independent Poisson streams,…

### Performance Bounds in Queueing Networks

- Computer Science
- 2011

This paper surveys computable performance bounds for sequencing and scheduling control in open networks to minimize the long run average number of customers, or a weighted average over different customer classes.

### Delay and Power-Optimal Control in Multi-Class Queueing Systems

- Computer Science, MathematicsArXiv
- 2011

The proposed unified framework provides a new set of tools for stochastic optimization and control over multi-class queueing systems with time average constraints and implement weighted priority policies in every busy period.

### Solving convex optimization with side constraints in a multi-class queue by adaptive $$c\mu $$cμ rule

- Computer ScienceQueueing Syst. Theory Appl.
- 2014

Combining the achievable region approach in queueing systems and the Lyapunov drift theory suitable to optimize renewal systems with time-average constraints, this paper shows that convex optimization problems can be solved by variants of adaptive $$c\mu $$cμ rules.

### Threshold control policies for heterogeneous server systems

- Mathematics, Computer ScienceMath. Methods Oper. Res.
- 2002

It is shown that any optimal, nonpreemptive policy is of threshold type, i.e., it assigns a customer to server Si, if this server is the fastest server available and the number of customers in the queue is mi or more.

### A linear programming approach to stability, optimisationand performance analysis for Markovian multiclassqueueing networks

- BusinessAnn. Oper. Res.
- 1999

A primal‐dual approachexploits the fact that the system satisfies (approximate) conservation laws to yield perform-ance guarantees for a natural index‐based scheduling heuristic, and is able to analyse the performance of an arbitrary priority policy.

## References

SHOWING 1-10 OF 66 REFERENCES

### Optimization of Polling Systems and Dynamic Vehicle Routing Problems on Networks

- Computer Science
- 1993

The proposed nonlinear programming relaxations provide strong lower bounds to the optimal cost for all static policies and also obtain lower bounds for dynamic policies as well, which are primarily useful under light traffic conditions and/or small switch-over times.

### Optimization of multiclass queuing networks: polyhedral and nonlinear characterizations of achievable performance

- Computer Science
- 1992

Using a powerful technique involving quadratic or higher order potential functions, methods for deriving polyhedral and nonlinear sets that contain the set of achievable response times under stable and preemptive scheduling policies are proposed and found that the first order approximation of the method is at least as good as simulation-based existing methods.

### Performance bounds for queueing networks and scheduling policies

- BusinessProceedings of 32nd IEEE Conference on Decision and Control
- 1993

A new technique for obtaining upper and lower bounds on the performance of Markovian queueing networks and scheduling policies is introduced, and analytic bounds which improve upon Kingman's bound (1970) for E/sub 2//M/1 queues are obtained.

### Characterization and Optimization of Achievable Performance in General Queueing Systems

- MathematicsOper. Res.
- 1988

The performance space, the set of performance vectors which are achievable under some preemptive work conserving rule, is a polyhedron described by 2J-1 inequalities, which allows for efficient procedures to minimize any separable convex function of the performance vector.

### Optimal control and scheduling in a multiclass queueing network: results and conjectures

- Business, Mathematics29th IEEE Conference on Decision and Control
- 1990

Dynamic scheduling and control in queuing networks are addressed, based on stochastic intensity representation of point processes, to minimize a discounted cost function over an infinite time horizon.

### Performance Bounds for Scheduling Queueing Networks

- Business
- 1992

The goal of this paper is to assess the improvement in performance that might' be achieved by optimally scheduling a multiclass open queueing network. A stochastic process is defined whose…

### Multiclass Queueing Systems: Polymatroidal Structure and Optimal Scheduling Control

- MathematicsOper. Res.
- 1992

A formal definition of strong conservation laws is developed, and it is shown that as a necessary consequence of these strong conservation Laws, the state space of the performance vector is a (base of a) polymatroid.

### A Characterization of Waiting Time Performance Realizable by Single-Server Queues

- BusinessOper. Res.
- 1980

Making use of a well known conservation law, this work proves a necessary and sufficient condition for the existence of a scheduling strategy that achieves the desired performance.

### Restless Bandits , Linear Programming Relaxations and a Primal-Dual Heuristic

- Computer Science
- 1994

A series of n increasingly stronger linear programming relaxations are introduced, the last of which is exact and corresponds to the formulation of the problem as a Markov decision process that has exponential size, while other relaxations provide bounds and are efficiently solvable.

### Branching bandits and Klimov's problem: achievable region and side constraints

- Computer ScienceProceedings of 1994 33rd IEEE Conference on Decision and Control
- 1994

The authors show that optimal policies can be obtained by randomizing between L+1 strict priority policies that can be found efficiently (in polynomial time) using linear programming techniques.