15-382 C OLLECTIVE I NTELLIGENCE – S18 L ECTURE 21: S WARM I NTELLIGENCE 7 / A NT C OLONY O PTIMIZATION 3 I NSTRUCTOR : G IANNI A. D I C ARO
T YPES OF PROBLEM SCENARIOS Offline: Full problem description and all data are accessible Online: Problem description / data become accessible sequentially, individually or in a bulk, while dealing with the problem. A model of data arrival can be available or not Stationary: Problem parameters / structure do not change over time (fixed), or keep being the same, on average Dynamic: Problem parameters / structure do change over time and do not follow any globally stationary pattern. A model of changes can be available or not 2
C OMPUTING / C ONTROL ARCHITECTURES Distributed problem Centralized problem § Data / code reside in physically § All data / code reside in the same different computing units computing unit / system § No shared memory: potentially need § Shared memory for communication channels § Shared time clock § Delays in information / code § Global system state available propagation § Updated global system state is not instantaneously available 3
C ONTROL ARCHITECTURES IN M ULTI -A GENT SYSTEMS Centralized MA control Distributed MA control § A central unit gathers data and § Control elements (controllers) are distributed perform decision-making for the throughout the system agents sending controls back § Final decision-making can be done in a centralized or decentralized manner, or in- between Decentralized MA control § Each peer unit / agent makes local autonomous decisions towards its individual (global) goals § Peers might interact with each other and share information or provide service to other peers § Emerging computation / behaviors § Local hierarchies can be present 4
M ORE ON CONTROL ARCHITECTURES Distributed Control System 15781 Fall 2016: Lecture 18 5
P ROBLEM TYPES : SO FAR AND NEXT ONES Ø The optimization problem we have considered so far for ACO are: § Centralized § Can be solved according to any centralized or distributed control scheme of choice § Parallelization / Multi-Threading is quite immediate § Stationary Ø Is the pheromone model (definition of stigmeric variables) the “same” for all types of centralized / stationary optimization problem? Ø ? What about problems that are: § Distributed § Dynamic § They seem “closer ” to ant colonies, by the way Ø Ant colonies perform automatic division of labor ( diverging stigmergy ) à 15781 Fall 2016: Lecture 18 6 Task Allocation Problems
D YNAMIC SCENARIO : D ATA ROUTING IN NETWORKS Ø Routing problems : move an item (object, vehicle, data, agent) from one start location to one final destination or to a set of other locations using some interconnection infrastructure (a network) Ø A number of constraints may need to be considered regarding: capacity of the vehicle and/or the infrastructure, time/length budget, ordering for visiting the locations, etc. Ø Goal: optimize the route followed to deliver/move the item TSP: routing of an agent over a set of predefined locations, minimize § traveled length VRP (Vehicle Routing Problem): routing of a vehicle over predefined locations § (to serve customers) subject to capacity and time constraints Network routing: routing of data across a telecommunication network § infrastructure (wired or wireless) o Distributed o Online 15781 Fall 2016: Lecture 18 7 o Constraints defined by the network technology
D ATA ROUTING IN NETWORKS § Decision system to forward data packets from origin to destination nodes § Unicast, multicast, broadcast, anycast § Optimization of network-wide metrics § Best-effort / QoS: Throughput, end-to-end delay, losses, energy, … § Constraints imposed by the used transmission technology § Wired (point-to-point, bus, optical), wireless (omni, LoS, DCF), ... § Constraints imposed by the hardware used for packet forwarding § Specialized routers, workstations, laptops, mobiles . . . 15781 Fall 2016: Lecture 18 8
N ETWORK ROUTING § Routing is the core component of the network layer in the OSI stack § The routing protocol specifies how routers communicate with each other to disseminate information useful for routing § The routing algorithm uses this information to build a routing table (routing information database) and implements route selection 15781 Fall 2016: Lecture 18 9
N ETWORK ROUTING : M ETA -P ROTOCOL At each network node / router: 1. Constant acquisition and organization of information concerning the local state : information on the local traffic flows and on the status of the locally available resources (.e.g., neighbors, connection capacity) 2. Build a view of the global network state , possibly by exchanging local state information with the other nodes 3. Use of the global view to set up the values of the local routing table and define the local routing policy aiming to optimize some measure of network performance 4. Forward user traffic according to the defined routing policy 5. Asynchronously and concurrently with the other nodes repeat the previous activities over time 10
L INK - STATE VS . D ISTANCE -V ECTOR ALGORITHMS § Each node maintains a complete Each node 𝑜 maintains a vector 𝐸𝑗𝑡𝑢 ( 𝑗,𝑒 ) of § description of the network distance estimates for reaching each network (hierarchical, dynamic) destination 𝑒 through each of it links 𝑗 § Flooding of link-state information § Distributed (asynchronous) DP for computing § Each node computes shortest paths optimal routes (e.g., Dijkstra-like algorithm) using Node 𝑜 = state of the DP algorithm, value (𝑜) § its local network description = distance to each possible destination 11
R OUTING IN M OBILE A D H OC N ETWORKS (MANET S ) Infrastructure-less network of mobile devices connected wirelessly § § Because of individual node mobility, wireless connection links can keep changing (breaking and creating) à Continual self-reconfiguring § Data traffic is forwarded in a multi-hop way , with each node acting as a source and a router at the same time § Wireless shared medium: collisions, interference, … § Heartbeat: each node must keep advertising its presence, in order to let neighbor nodes aware of its presence (and possible use as a router) 12
A NT H OC N ET : ACO FOR ROUTING IN MANET S § Goal: Adaptively, build routing tables for data forwarding § Routing tables have an entry for each known destination 𝑒 and each known current neighbor 𝑜 § “Known”: things keep changing! 𝑅 𝑜𝑒 is a pheromone variable, it represents the § estimated goodness/quality of routing decision 𝑜 for reaching destination 𝑒 (i.e., forwarding to neighbor 𝑜 a data packet with destination 𝑒 ) § Data packets are forwarded probabilistically, proportionally to the value of pheromone variables § Basic ACO idea: Ant agents are generated at each node to sample (solution) paths to known destinations, pheromone variables are used to guide path search, and are updated as a result of the found paths (ant agents need to physically retrace, hop by hop, the travelled path from source to destination) Challenge: Ideally, the more ant agents are generated, the more paths are sampled, § and the better the pheromone values can be estimated. Unfortunately, generating many ants creates traffic congestion and make the network unusable à Finding a good tradeoff between path sampling and bandwidth usage is necessary 13
A NT H OC N ET : ACO FOR ROUTING IN MANET S What exactly pheromone variables represent? The quality of a path in a MANET § depends on many aspects: associated end-to-end delay, but also the number of hops, interference at each hop, degree of mobility of path nodes 𝑅 = Low delays AND High expected stability and reliability of the path § 𝑅 -. = 𝐺 ( SNR, end-to-end delay, hops, mobility, … ) § Hybrid algorithm: § Reactive: paths are only setup at the starting of a session § ( ~ controlled flooding, high bandwidth usage) Proactive: during the course of a session paths are proactively improved § and extended with a low impact on bandwidth AntHocNet integrates two mechanisms for distributed adaptive learning: § Monte Carlo sampling and updating of full paths with ants (ACO) § Periodic exchange of routing information among the nodes, and local § incremental updating of the estimated pheromone values using a mechanism equivalent to Dynamic programming / Bellman-Ford 14
Recommend
More recommend
