|
Energy-Aware Opportunistic Charging and Energy
# C4 a. B( G; B7 ADistribution for Sustainable Vehicular Edge and Fog
0 G( t8 K; w* C# c: ZNetworks 2 _% e6 ]1 x, t! V- @8 E! g/ M
6 |- e' t! g) J( ?( l% t
* Y }! J" P: R8 p& GThe fast-growing popularity of electric vehicles 6 T* A }2 k, z& `- Q; y% p. `
(EVs) poses complex challenges for the existing power grid ' d' o4 L2 K( [: Q* A; ]' D4 Y( Q1 P
infrastructure to meet the high demands at peak charging hours. 3 ]+ T. e+ X$ j+ J7 Y E6 d) n( R! e
Discovering and transferring energy amongst EVs in mobile
; `% U( }- ]* H7 d2 d0 Z6 Z( ]; Avehicular edges and fogs is expected to be an effective solution for
( f i* }& z* q- L# r3 R1 n, Nbringing energy closer to where the demand is and improving the ( |& F2 t' g5 D/ Z! a' \
scalability and flexibility compared to traditional charging 8 x$ G6 f/ s3 m( ^& w, N
solutions. In this paper, we propose a fully-distributed energy
1 s* y6 T) u: q8 Z8 {3 Xaware opportunistic charging approach which enables distributed . A1 g- h. [: n3 k% E
multi-layer adaptive edge cloud platform for sustainable mobile " L- U. Q9 K$ ?2 b |
autonomous vehicular edges which host dynamic on-demand ! F4 m) S6 Q5 ^0 g
virtual edge containers of on-demand services. We introduce a 7 G8 o$ z* w3 {+ t* o
novel Reinforcement Learning (Q-learning) based SmartCharge
& J- d. K- P. ?algorithm formulated as a finite Markov Decision Process. We , u7 t0 i! `( Q* O
define multiple edge energy states, transitions and possible actions ( ]# V1 g2 g( Q' X
of edge nodes in dynamic complex network environments which
, V( O4 m0 R! n3 d9 G$ J! [4 ^are adaptively resolved by multilayer real-time multidimensional
! Z8 E/ z* k! N& q% ?2 Q9 Ypredictive analytics. This allows SmartCharge edge nodes to more
2 h( s6 P& C) U/ H+ H. Qaccurately capture, predict and adapt to dynamic spatial-temporal # P. z( Z" u* E ?' G
energy supply and demand as well as mobility patterns when
: Y" @# N" f7 s4 Lenergy peaks are expected. More specifically, SmartCharge edge 8 H- W: C- K# c) H" |$ Z
nodes are able to autonomously and collaboratively understand 6 e9 R- c& e3 @+ ?2 B
when (how soon) and where the geo-temporal peaks are expected
7 I- N0 V3 G n) u; gto happen, thus enable better local prediction and more accurate : }5 |% d$ s3 r0 u- R+ c! e5 d
global distribution of energy resources. We provide multi-criteria
3 A" I6 D' H' eevaluation of SmartCharge against competitive protocols over , s3 y! K* M8 D% h$ P( M9 d5 g
real-world San Francisco Cab mobility traces and in the presence : u, y7 k4 D/ }' B! C; N: q
of real-world users’ energy interest traces driven by Foursquare
$ g: o. W4 @8 g" D6 aSan Francisco dataset. We show that SmartCharge successfully + i& I3 f {) k$ ?
predicts and mitigates congestion in peak charging hours, reduces # @$ c3 U0 W- |1 a( q! R$ g+ a
the waiting time between vehicles sending energy demand requests
: ^% C. X3 S- J# P+ Q4 Yand being successfully charged as well as significantly reduces the % G8 b5 F) W; @3 F
total number of vehicles in need of energy.
& V6 N2 @1 P$ e) O2 L; Z3 d
/ F+ c% Y' o9 V
* m& N% y5 l& C$ v; u! }" t! j: P( c2 p" s; D# d
1 \1 x- I) `6 d3 u# \* _
| 
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
