|
Energy-Aware Opportunistic Charging and Energy
3 y; K% r6 b4 f8 V5 QDistribution for Sustainable Vehicular Edge and Fog # H- H8 A+ D0 Z2 l# M' |6 F
Networks : a1 ~7 @) k* [$ k4 f. ] x
$ X( ~ [3 u+ j0 e
9 O8 j2 F8 {% O7 b- c
The fast-growing popularity of electric vehicles
9 f& ]/ r3 `. h" g; Z: ` X( y(EVs) poses complex challenges for the existing power grid 4 ?4 }1 B' e4 K1 u- M1 P N
infrastructure to meet the high demands at peak charging hours. ) F4 B: i O+ [" _, b
Discovering and transferring energy amongst EVs in mobile + m2 f3 B5 ?6 R& v! ?8 y& C O
vehicular edges and fogs is expected to be an effective solution for
4 |& L$ ~6 O+ R0 h& Y+ G1 \: ybringing energy closer to where the demand is and improving the 6 B+ A0 {* Y- t: F/ I. S
scalability and flexibility compared to traditional charging
6 F) ?8 K8 d5 esolutions. In this paper, we propose a fully-distributed energy
J" O$ d) c& U5 h* Q) x2 k8 E' Daware opportunistic charging approach which enables distributed
0 x5 a# T# ^$ H9 Mmulti-layer adaptive edge cloud platform for sustainable mobile ' m* v! p) Q* d, \
autonomous vehicular edges which host dynamic on-demand ( t" v0 y8 z8 [* \$ L; u$ [
virtual edge containers of on-demand services. We introduce a
2 g3 C3 M5 ?* [$ o9 Knovel Reinforcement Learning (Q-learning) based SmartCharge $ t: W9 T& G* E( T& {
algorithm formulated as a finite Markov Decision Process. We 0 T' I6 X$ b& N6 z
define multiple edge energy states, transitions and possible actions 7 b' a8 a' X d- R( b8 h7 ~
of edge nodes in dynamic complex network environments which
) E9 V! U7 i- M) rare adaptively resolved by multilayer real-time multidimensional 8 F& h- v- |& T
predictive analytics. This allows SmartCharge edge nodes to more & h3 K% N* T( e3 s, E
accurately capture, predict and adapt to dynamic spatial-temporal
0 r% V( K( a7 o6 Q( u& W. {% n: denergy supply and demand as well as mobility patterns when : o% N) Q) o$ k: l
energy peaks are expected. More specifically, SmartCharge edge z) p, f/ M2 p1 T& |
nodes are able to autonomously and collaboratively understand
9 ?9 r7 L. [9 {1 i- J/ Mwhen (how soon) and where the geo-temporal peaks are expected
. n' ~. c2 P* h" [- ?& Q6 G9 {to happen, thus enable better local prediction and more accurate
5 z9 h$ F& l0 h5 K& ^global distribution of energy resources. We provide multi-criteria
. E8 _7 S; g. @1 C2 _evaluation of SmartCharge against competitive protocols over 4 J" y3 P! K/ T$ z8 B
real-world San Francisco Cab mobility traces and in the presence
7 A" ?, ]. j; xof real-world users’ energy interest traces driven by Foursquare 1 c% y% B; G9 H1 j( ^& V# o- v+ ~
San Francisco dataset. We show that SmartCharge successfully
2 @! ^& B \. @0 c1 j+ ^predicts and mitigates congestion in peak charging hours, reduces 6 R1 r7 P4 z7 l" z
the waiting time between vehicles sending energy demand requests 7 j9 H/ a5 n6 x4 K# d D; D8 Q
and being successfully charged as well as significantly reduces the ; ~7 W" r. i2 t" S6 X- H
total number of vehicles in need of energy.
, @$ M$ k, d! `( @1 U% n
' Q) n7 n. D) v; l8 }2 ?5 x% ^
- t y; Z0 j3 ~" }1 m' G' e8 t" y8 \3 ]. G1 y' W) {+ I- p
) ]8 i& D3 h: g) P% X5 R. J
| 
IP卡
狗仔卡
发表于 2020-11-9 15:10
转播
淘帖
分享
收藏
支持
反对
微信
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
抢沙发
显身卡
