|
Energy-Aware Opportunistic Charging and Energy 0 e* W/ ^3 r9 I |$ K$ r- R
Distribution for Sustainable Vehicular Edge and Fog / ]; `0 A3 h* T6 ?5 ?. k$ w4 m
Networks 3 m- S- g( h9 \
& A* c- r& C+ U7 T' ]
8 p+ x. N" W; v; t: s: L+ KThe fast-growing popularity of electric vehicles
2 p9 ~ N3 p3 w5 l(EVs) poses complex challenges for the existing power grid
4 k8 w5 _7 s/ _: r& D& Jinfrastructure to meet the high demands at peak charging hours.
- C. _. F* P- T+ s( W, `Discovering and transferring energy amongst EVs in mobile 4 D* Q9 V0 L1 Z1 d# p( Q1 G
vehicular edges and fogs is expected to be an effective solution for
4 K0 g3 N6 s7 P0 [2 w' ebringing energy closer to where the demand is and improving the
+ n+ R1 B- |6 sscalability and flexibility compared to traditional charging * A: O! V" M9 ^
solutions. In this paper, we propose a fully-distributed energy
& ~% W' z6 {* U# [aware opportunistic charging approach which enables distributed
. I, H$ j6 F% F! g0 Omulti-layer adaptive edge cloud platform for sustainable mobile
8 B" b/ O, k# ]2 R1 o6 h3 f& {autonomous vehicular edges which host dynamic on-demand
4 I& m: O; A3 N# `2 Qvirtual edge containers of on-demand services. We introduce a
: L+ D% @9 ^, W; g1 {0 O5 wnovel Reinforcement Learning (Q-learning) based SmartCharge
0 n& x% }0 _. }algorithm formulated as a finite Markov Decision Process. We ) e* k7 \5 G/ m' s2 D7 F7 X- ^
define multiple edge energy states, transitions and possible actions
* [: W b2 l! |! Uof edge nodes in dynamic complex network environments which
( t; _+ I3 O% Fare adaptively resolved by multilayer real-time multidimensional 4 f) S# p/ w8 O% @
predictive analytics. This allows SmartCharge edge nodes to more 0 A+ K$ E3 b0 U7 j% \$ y: ^0 s
accurately capture, predict and adapt to dynamic spatial-temporal
5 b$ p6 u+ Q/ p! {* h0 C$ Renergy supply and demand as well as mobility patterns when 7 `) _9 W j; m
energy peaks are expected. More specifically, SmartCharge edge
* J+ V k( C: N! K1 ~7 Dnodes are able to autonomously and collaboratively understand
2 X( v- L( H& x f2 s5 n/ awhen (how soon) and where the geo-temporal peaks are expected ) S7 z3 p z- r) I9 d
to happen, thus enable better local prediction and more accurate
1 D2 ?( \" \1 M, v" Aglobal distribution of energy resources. We provide multi-criteria
* P. z* K2 x% X" y! ?evaluation of SmartCharge against competitive protocols over
3 q3 o) z( x$ k4 i7 `real-world San Francisco Cab mobility traces and in the presence
* Q, @' J" Z6 {; Yof real-world users’ energy interest traces driven by Foursquare
5 d; [" e* ]# N) ^" G n4 BSan Francisco dataset. We show that SmartCharge successfully & q# B& q6 f# p$ s5 g5 s
predicts and mitigates congestion in peak charging hours, reduces
. ]! |% x! r! e7 Y# N" x; w. {the waiting time between vehicles sending energy demand requests
" G, P- _$ |$ _/ Q6 w( Mand being successfully charged as well as significantly reduces the
1 A3 m$ E; D! G* s) @. atotal number of vehicles in need of energy. 1 U+ g- z( D. H/ }' W
" W( X: k6 u) E2 @% k; i: _; v4 G
2 l- u4 U- D/ D; r3 r
, {/ I& n/ F+ H7 e1 p4 U, m
# ^- c! n; v' \9 B
| 
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
