Application Design Aiming to Minimize Drivers’ Trip Duration through Intermediate Charging at Public Station Deployed in Smart Citiesâ€
Round 1
Reviewer 1 Report
The paper present a framework based on an algorithm allowing the management of charging plans for electric vehicles traveling on the road to their destination, in order to minimize the duration of the drivers’ journey including waiting and charging times. In order to show the effectiveness of the proposed approach a comparison with one algorithm within the existing literature is performed and discussed.
Although the paper is well written and adequately structured, in the final paper the authors should address the following issues:
The Authors should enrich the bibliographic survey in order to provide more insights to readers and they should better highlight the contribution of the paper related to the existing literature. Section 2: EVPSS is not defined in the paper. Section 2.2, row 167. The Authors state: “Each EVPSS has a multi-service queue with identical servers of Number Ns, each operating with an exponential service rate u”. Is the multi servers queue limited size buffer or an infinite size buffer? In the latter case, the Authors have to consider that although the closest EVPSS to EV could not accept the charging request. Section 2.2, row 169. The Authors state “The input data is the EV requiring a power load…”. The Authors should consider also the residual State of Charge of the EV (i.e. the required energy of the EV) in order to perform an effective management of the charging request. Section 2.3: Does the Authors consider any priority mechanisms in charging requests of the EV user? Section 2.3, row 182-183. The Authors state: “After traversing all the CSs, the algorithm returns the CS to which the VE driver will spend a minimum trip time, including the charging time, line 18”. Is the queue waiting time included in the minimum trip time computed by the algorithm? Section 3. Have the Authors performed Monte Carlo simulations in order to perform the performance evaluation analysis?Section 3. The Authors compared the performance of the algorithms by increasing and decreasing of the charging stations. What about if the number of EVs in the network and/or at the charging stations is increased?
Author Response
Point 1: The Authors should enrich the bibliographic survey in order to provide more insights to readers and they should better highlight the contribution of the paper related to the existing literature.
Response 1: We've added some new paper that all touch on the same subject as ours and we presented their work in the introduction.
Point 2: Section 2: EVPSS is not defined in the paper.
Response 2: electric vehicle public supply station (EVPSS) is defined in row 94-95
Point 3: Section 2.2, row 167. The Authors state: “Each EVPSS has a multi-service queue with identical servers of Number Ns, each operating with an exponential service rate u”. Is the multi servers queue limited size buffer or an infinite size buffer? In the latter case, the Authors have to consider that although the closest EVPSS to EV could not accept the charging request.
Response 3: The queuing size in the EVPASS isn't taken into account by our decision algorithm and we considered that the stations having enough capacity to accommodate the EVs. Otherwise, we will be turned to another optimization problem depends areas occupied by the EVSPP and parking closest to it, also the type and size of EVs.
Point 4: Section 2.2, row 169. The Authors state “The input data is the EV requiring a power load…”. The Authors should consider also the residual State of Charge of the EV (i.e. the required energy of the EV) in order to perform an effective management of the charging request.
Response 4: To compare the two ideas one looking for the station that has the minimum waiting time to charge and the other looking to minimize the trip time, including charging time, we assumed that each EVPSS at a fixed charge rate equals u and in simulation, we considered that each EVPSS charging slot has a charging power equal to 3 EVs per hour (u = 3).
Point 5: Section 2.3: Does the Authors consider any priority mechanisms in charging requests of the EV user?
Response 5: No priority mechanism has been adopted by the proposed system except first-come-first-serve (FCFS) policy used in EVPSSs to distribute EVs to charging sockets.
Point 6: Section 2.3, row 182-183. The Authors state: “After traversing all the CSs, the algorithm returns the CS to which the VE driver will spend a minimum trip time, including the charging time, line 18”. Is the queue waiting time included in the minimum trip time computed by the algorithm?
Response 6: The queue waiting time is included in the minimum trip time that estimated by the algorithm and this expressed by the instruction " TTrip = Tarr + Twait + Tcs,evd " in row 12 of the algorithm , Twait represents the waiting time
Point 7: Section 3. Have the Authors performed Monte Carlo simulations in order to perform the performance evaluation analysis?
Response 7: In row 222, we reported that electric vehicles are distributed uniformly random over the adopted geographical area using the Java Random class. The latter contains methods such as nextInt (), nextDouble (), etc., used to generate the random values
Point 8: Section 3. The Authors compared the performance of the algorithms by increasing and decreasing of the charging stations. What about if the number of EVs in the network and/or at the charging stations is increased?
Response 8: The performance of the algorithms during the evaluation if we fix the number EVs and increase the number of EVPSS is the same as when we fix the number of EVPSS and decrease the number of EVs.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper focuses on an interesting topic - smart charging. However, there are some places need to be improved.
How the EV itineraries generated? How to judge the charging stations can satisfy the EV charging load? including the rated charging power, the working hours... I noticed that there is a constraint that the travel speed is fixed to 16 m/s, which means road congestion is neglected. But, how the geography road influence the results? Is there an energy consumption rate for the EV? What's the charging rate in the algorithm? What's the charging behavior? Algorithm 1 and Algorithm 2, how to make sure they have the same initial values, e.g. initial SOC, trip itineraries....Author Response
Point 1: How the EV itineraries generated?
Response 1: In row 222, we reported that electric vehicles are distributed randomly over the adopted geographical area using the Java Random class.The latter contains methods such as nextInt (), nextDouble (), etc., used to generate the random values that affected to the EV itineraries.
Point 2: How to judge the charging stations can satisfy the EV charging load?
Response 2: It is assumed that all EVPSS linked to the electricity network continuously to satisfy the needs of all EV types. thus, at the beginning of each run of the algorithm by GA, it only load the active EVPSSs to the list, row 1 of the algorithm, to avoid any recommendation from an EVPSS that is out of order or discontinuous of electricity network.
Point 3: Including the rated charging power, the working hours... I noticed that there is a constraint that the travel speed is fixed to 16 m/s, which means road congestion is neglected. But, how the geography road influence the results? Is there an energy consumption rate for the EV?
Response 3: To have a comparison between two algorithms we've assumed, in the simulation, that the EVs maximum speed cannot exceed 16 m/s in the cities without taking account the congestion.
Point 4: What's the charging rate in the algorithm? What's the charging behavior?
Response 4: .in order to compare between two ideas one looking for the station that has the minimum waiting time to charge and the other looking to minimize the trip time, including charging time, we assumed that each EVPSS at a fixed charge rate equals u and in simulation, we considered that each EVPSS charging slot has a charging power equal to 3 EVs per hour (u = 3), table 2 row 228
Point 5: consider Algorithm 1 and Algorithm 2, how to make sure they have the same initial values, e.g. initial SOC, trip itineraries....
Response 5: In the simulation that we have programmed, we put two servers, global aggregator, one for algorithm 1 and the other for algorithm 2. Thus, the system affects to each new EV created their current itineraries and their trip itineraries in a random way. After, we redefine the clone method of the language java which returns us a copy of the new EV created. This allows us to obtain two EVs identical; the charging request of one of these EVs is treated by GA1 of algorithm1 and the other treated by GA2 of algorithm 2.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Interesting work. The number of the simulated vehicle could be increased in future work.
