An Effective Resource Allocation and Revenue Generation for Rental Vehicles

There are various studies on seat selection in the traveling mode in which the majority were on seat bookings in the flights (airlines). Among them, specific studies focus on per descriptions of references listed. From Guerriero and Olivito [1], the discussion is on applying one method such as dynamic programming and logistic operator in the car rental zone is framed as a policy. In various scenarios, the specific criteria are addressed in overcoming the challenges faced during this scenario. In regard to Haensel et al. [2], the usage of capacity controls and fleet distribution forms a novel method called stochastic program at rental stations. The result is analyzed over a small network management over a set of scenarios.

In the view of the study [3], the analysis is done from 2019 to 2027 that involves the evolution of technology in which self-driving cars and autonomous features would dominate as per prediction and usage of vehicles. This leads to a reduction in rental prices. The graph also demonstrated showing technology usage and price drops. From the study of Oliveira et al. [4], the profit computed using pricing and capacity integration and biased random key generation method against a hierarchical approach is compared and discussed. It makes use of initial prices, fitness calculation, decomposition, and relaxation concepts, and even uses constraints. From Yang et al. [5], it compares a proposed model called the hybrid approach that consists of Holt-Winters as linear and backpropagation as nonlinear models. The bias errors, effectiveness, and feasibility are addressed, and are satisfied with the results obtained.

In the context of Fiig et al. [6], discussion on methods such as single leg model and multiple airline network models experienced that increase or decrease price elasticity parameter bias. This would result in a revenue reduction of 1% or 4%. This work is oriented towards airline revenue.

From Wang et al. [7], the invented approach implemented in two modes which first mode works based on decision tree induction for decision making, and the second mode finishes analysis using Shapley additive prediction (SHAP) using important influencing factors (such as the price of the ticket, preferences of the user, and duration of journey) that results in better interpretation. In the regard of Rouncivell et al. [8], the demonstration of bi-variant analysis for seat selection results in either of negatively (price of seats) or positive (reputation and frequency of takeoffs) correlation. The efficiency of this study depends on past experience of seat selection, fares, frequency of take-offs and results of ancillary returns.

From the study of Ivanov [9], demonstrates the logic of to whom seat selection is to be disabled, and to whom, the seat selection to be enabled could be evaluated by the seat selection algorithm. In this, the lock is imposed on selected stations, before the starting station, and after the target station, the lock is released that allows to book of the tickets if there are seats. The study of Boyar and Larsen [10], demonstrates an algorithm called FFBF which stands for First Fit Best Fit approach (based on FCFS), allows to book for long length duration stations than short length. The internal algorithm used to confirm the seats is a randomized method. From the study [11], it demonstrates the various levels of booking tickets using VIP, ladies, impaired level, and old citizen. The preference is to be given to the passengers who are of type single women, old couples, newly married couples, children, and handicapped people but there is no existence of selection of seats before tatkal and at the tatkal time.

From the study [12], it elevates how to log into the railway system, search for a train, enter passenger details, and make payment. One payment is made, and tickets to be confirmed if there are vacancies are there. The description is elaborated using ER, DFD, activity, and sequence diagrams. The study [13] demonstrates the Java code for movie ticket booking, in which seat is the two-dimensional array that keeps track of filled and not filled seats. Deshpande et al. [14] demonstrate the car rental app for users who are traveling to high rates places, and many other characteristics are taken into focus such as vehicle type, location, history, and rental patterns. The analysis of the application allows to use its services effectively.

From the study [15], it shows various policies set for customers, drivers, and owners of the app, and imposes feedback on those to evaluate, and monitor the progress of the app, and its services. Chen et al. [16] demonstrate student choice of seating environment, which helps the student to involve in the course delivery and activities. The environment builds student learning experience, and strategies, and facilities are available that nurture all-round fusion gamified scenarios. This environment addresses and overcomes various competencies, and provides a holistic atmosphere that is suitable to do practicals, activities, and multi-modalities.

From Losonczy-Marshall and Marshall [17], demonstration of the bifurcation of students’ mindset and their performance, other factors would be influenced by the selection of seats. The student seat selection is assessed by a set of questionnaires. According to their answers, these questionnaires were grouped into five categories. The seat sitting may be among first, back, sides, middle, corner, etc. Zhao and Zhao [18] demonstrate an increase in earnings especially in the HSR route by booking tickets using nonlinear programming. The various factors are taken into consideration such as preferences, specific timing, number of stops, and different segments of the compartments, etc. This exercise and practice would increase the revenue management of railways. Ko et al. [19] demonstrate re-assigning the seats when an aircraft was canceled and assigned another aircraft for traveling in which seat allotment to be taken place. The mathematical-based allocation is used which assigns intelligently when there is grouping of seats are there. In the airways, the seat selection taken place in advance as per availability and customer choice. From the study [20], it demonstrates the variety of layouts that event planners are supposed to apply for the events and may experience the customer feedback. It leads to innovative styles of arrangements of the seating patterns and allows registering the customers by group or individual as per their preferences.

Milne et al. [21] demonstrate a staged procedure that takes the passengers based on their luggage in the first phase, and weighted slack times to be maximized using a mixed integer programming concept. The outputs demonstrated using this approach would be better using this proposed model. Zhou et al. [22] demonstrate the increase of revenue using the logit model and annealing approach by applying various pricing based on passenger demand divisions. The results obtained would guarantee better service and more revenue than present models on the route from Beijing to Shanghai in China. Hales and García [23] demonstrate the seating arrangement of political parties in assembly sessions and in congress meetings using the mathematical approach in which the heuristic model works efficiently and produces good results than other existing methods.

From Fabro [24], demonstration of strategies, and policies for booking seats for the restaurant system. During which, double booking is to be avoided. Have to welcome and assign the seat during their arrival who already booked the table or seat which is trending in the restaurants. Deng et al. [25] present the use of a mathematical optimization model over the Wuhan route to increase revenue. The imposition of dynamic pricing rates based on the demand of passengers, and price fixes more over short-distance trains than long-distance trains, and also based on the point where more passengers are there, again dynamic pricing and seat selection to be applied. Weicker [26] demonstrates minimizing the distance between the trains when there is transferring the seats, uses an optimization model, and handles NP-hard problems present in the scenarios.

Ricky et al. [27] demonstrate challenges, and issues raised in ranking algorithms such as repetitive data, and multi-dimensional classes. It also addresses pair-wise, types of two bipartite methods, and their working themes. Regarding Efimov [28], it demonstrates three optimal ranking approaches such as pair-wise, point-wise, and list-wise, and discusses their working procedure along with benefits and difficulties. Hrushikesava et al. [29] provide demonstrations on specific places based on the mass visit that, automatically raises the feedback of the customer. The accumulation of reviews of many such visitors is computed and makes that place more popular. The higher number of visitors of the place would mark a high rank for that place. This study did not describe optimality but focused on ranking of the places. From Bulum et al. [30], the focus done on revenue generation by using ticket booking system in which buckets assigned for each booking, proved some raise tickets purchase. This study went on rails, as shift from airways. There is still significant increase of the revenue to be expected, which also motivates for proposed system to design. Most of these studies have not dealt with the optimality of the service and are not aimed at revenue generation for the rental vehicles.

The proposed approach derives its capability over accuracy, performance, and cost metrics. In Table 3, the purpose and advantage of each module is highlighted. The comparison of the hybrid optimal (novel) approach against existing approaches is demonstrated in Tables 4-8.

Table 3. The advantages of listed modules

The analysis of existing applications if they would import this kind of efficient model, they can forecast or predict future days income to auto-calendar notification feature enabled. The future scope would be environmental consequences which may also drop the income that needs to be incorporated in this model.

The comparison of existing methods versus proposed method hybrid optimal approach are demonstrated in the below Table 4.

Table 4. Metrics comparison of hybrid optimal against existing

The accuracies of novel optimal against existing approaches are demonstrated in Table 5, and depicted in Figure 3. Accuracy is defined as correctly predicted instances by total instances. Among these approaches, novel optimal approach is observed having high accuracy than others.

Accuracy = Instances correctly computed / Total instance cases

Table 5. Accuracy of hybrid optimal against existing

The efficiency of novel optimal against existing approaches are demonstrated in Table 6, and depicted in Figure 3. Efficiency is defined as achieved performance to be looking good even though the number of instances is to be increased. Among these approaches, novel optimal approach is observed having high performance than others due to simple number of constraints, and their conjunctions.

Table 6. Performance of hybrid optimal against existing

3.png

Figure 3. Accuracies, performances of hybrid optimal system against existing approaches

The computation cost of novel optimal against existing approaches are demonstrated in Table 7, and depicted in Figure 4. The computation cost is defined as the amount spent to achieve the task to be done effectively. Among these approaches, novel optimal approach is observed having medium computational overhead which is accepted than others. The MSE of hybrid optimal against existing approaches are demonstrated in Table 8, and depicted in Figure 5. The MSE's of existing approaches, against hybrid optimal approach is observed having less impact which is extremely accepted than others.

Table 7. Cost of hybrid optimal against existing approaches

Table 8. Error rate of hybrid optimal against existing approaches

4.png

Figure 4. Cost of optimal against existing

5.png

Figure 5. Error rate (MSE) of hybrid optimal against existing approaches