|
Energy-Aware Opportunistic Charging and Energy
/ i* h; o* K7 N, EDistribution for Sustainable Vehicular Edge and Fog ; E6 M. M2 I" T1 _2 i4 s \5 A* M
Networks
+ v- Y) Y8 w, e: _; a% w6 w4 e% i3 S% K) o! L8 v0 p
- D4 L2 c+ _; IThe fast-growing popularity of electric vehicles
' ?: y! N/ e( y(EVs) poses complex challenges for the existing power grid * I* Y; ^, L. _
infrastructure to meet the high demands at peak charging hours.
* A$ n) Q/ }, c; L# fDiscovering and transferring energy amongst EVs in mobile . d1 a6 s( n& g5 a1 h$ s
vehicular edges and fogs is expected to be an effective solution for
7 l: V R4 ~, ]bringing energy closer to where the demand is and improving the & k i8 Q( a# J+ g7 L5 E
scalability and flexibility compared to traditional charging E5 i! S: C0 r9 F9 s0 k
solutions. In this paper, we propose a fully-distributed energy2 M- F Q5 L0 B' x: ^4 a. w, R
aware opportunistic charging approach which enables distributed
# d( x$ _0 i/ x4 w rmulti-layer adaptive edge cloud platform for sustainable mobile 3 A8 l! b, i9 I
autonomous vehicular edges which host dynamic on-demand
& S6 w" V/ s) Y% u6 ~virtual edge containers of on-demand services. We introduce a
$ t& g& D g# tnovel Reinforcement Learning (Q-learning) based SmartCharge + U8 j$ H3 Z) ?4 f" f
algorithm formulated as a finite Markov Decision Process. We * p7 t: p2 E( W. G' |) M3 Y. l+ G4 [# ?8 H
define multiple edge energy states, transitions and possible actions $ K$ B) m, ~* X# i0 X
of edge nodes in dynamic complex network environments which
& `8 f8 T$ C) {$ [are adaptively resolved by multilayer real-time multidimensional # K4 j2 Z2 W9 a; \* {- O, s/ m5 ]
predictive analytics. This allows SmartCharge edge nodes to more
8 b ^. g- r9 ?- _# r8 Haccurately capture, predict and adapt to dynamic spatial-temporal % J6 ^4 t+ Q, r# x1 U/ t6 D; S
energy supply and demand as well as mobility patterns when
' P: d" A$ ~' D3 |) @' B; menergy peaks are expected. More specifically, SmartCharge edge |- \$ t5 s+ C8 ~8 G A1 p
nodes are able to autonomously and collaboratively understand * r U' W) d% I. m
when (how soon) and where the geo-temporal peaks are expected
% Q9 d( [* R" f4 F) dto happen, thus enable better local prediction and more accurate
6 | A' ^* e7 k9 L2 }( R( Eglobal distribution of energy resources. We provide multi-criteria
8 x$ i; q6 F: z5 N$ U/ ]8 {$ ^evaluation of SmartCharge against competitive protocols over : w1 p, V9 ]( z5 e+ H! |) [0 t
real-world San Francisco Cab mobility traces and in the presence
$ G0 |2 m) C! A# v. [2 @of real-world users’ energy interest traces driven by Foursquare $ |* V$ p2 z% T `4 [9 u
San Francisco dataset. We show that SmartCharge successfully
" [8 R' z7 L! Y O6 b7 ]predicts and mitigates congestion in peak charging hours, reduces " ^- p* ?3 l- d4 |6 E
the waiting time between vehicles sending energy demand requests + x' F- L' \; Y& P7 ~+ c+ q: D9 P
and being successfully charged as well as significantly reduces the
( V7 J" s" c9 X: _total number of vehicles in need of energy. # ?) q w; X! [! v2 c7 S' b
# ?# a& ]4 {* B: R6 d9 x% m/ }" J
- n. {! E4 C, |2 n/ ?! d/ M% |+ G& h6 {9 a; C$ e* w
$ o: ^" {/ z- ` ` | 
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
