Scheduling Optimization in Wireless MESH Networks with Power Control and Rate Adaptation SECON 2006 Antonio Capone and Giuliana Carello Keon Jang
Outline Introduction Problem Formulation Solution Approach Numerical Result Conclusion
Introduction Wireless Mesh Networks (WMNs) have emerged recently as a new network architecture WMNs can partially replace the wired backbone network The backbone network is usually devised to provide an almost static resource assignment
Introduction(cont’d) Optimizing resource utilization rather than bandwidth or latency can ease and expediting admission control that need to estimate spare bandwidth Scheduling transmission with optimization of power and rate jointly can improve overall network performance
Approach Assuming TDMA, power control and rate control ability for each slot Find a scheduling that minimizes the number of used time slots to guarantee the required bandwidth Integer Linear Programming(ILP) and heuristics to improve computing time
Problem Formulation IP Formulation ▫Will be transformed to ILP Objective Function ▫Minimize the number of used slots
Problem Formulation(cont’d) Constraints ▫x k is one if at least one link is active in time slot k ▫Each device is active in at most one link in each slot
Problem Formulation(cont’d) Constraints ▫Traffic requirement of each link is fulfill ▫SINR is enough for transmission Traffic Requirement Minimum SINR Gain on link (i,j) Transmission Power
Problem Formulation(cont’d) Without considering power control Considering power control
Problem Formulation(cont’d) Without considering Transmission Rate Considering Transmission Rate
Problem Formulation(cont’d) Column Generation Model ▫In general, column generation means that not all variables will be considered explicitly; columns will be added to the problem “on the fly” 1: Solve Problem with initialConfiguration 2: While(adding variable improve solution(Pricing problem)) 3: Add variables 4: Solve Problem with current Configuration
Solution Approach Linearization After Linearization
Solution Approach Performance Optimization ▫Pricing problem is ILP(takes long time) ▫In column generation model, instead of solving pricing problem, use greedy algorithm to see if solution can be improved 1: Solve Problem with initialConfiguration 2: While(adding variable improve solution(Pricing problem)) 3: Add variables 4: Solve Problem with current Configuration
Numerical Results Small experiment that shows an effects of power control and transmission rate control
Numerical Results 8 randomly generated simulation for each size
Conclusion Formulated the joint scheduling, power control, and rate control problem in WMN Proposed an heuristic approach that gives an good solution over initial set