Multi-Objective Linear Optimization Problem for Strategic Planning of Shared Autonomous Vehicle Operation and Infrastructure Design

Abstract

This study proposes a unified optimization framework for strategic planning of shared autonomous vehicle (SAV) systems that explicitly and endogenously considers their operational aspects based on macroscopic dynamic traffic assignment. Specifically, the proposed model optimizes fleet size, road network design, and parking space allocation of an SAV system with optimized SAVs' dynamic routing with passenger pickup/delivery and ridesharing. It is formulated as a multi-objective optimization problem that simultaneously minimizes total travel time of travelers, total distance traveled by SAVs, total number of SAVs, and infrastructure construction cost; thus, both the user-side cost and the system-side cost are taken into account, and their trade-off relations can be explicitly investigated. Furthermore, the problem is formulated as a linear programming problem, making it easy to solve. By leveraging the linearity, we mathematically derive a useful property of the problem: introduction of ridesharing can weakly monotonically and simultaneously decrease the user-side cost and system-side cost. The proposed model is evaluated by applying it to actual travel records obtained from New York City taxi data.

Ridesharing and Crowdsourcing for Smart Cities: Technologies, Paradigms and Use Cases

“…The review has covered the various models and architectures for (dynamic) ridesharing and crowdsourcing. The paper has also given insights into the various techniques and approaches which have been utilized Network topology for on-demand ridesharing systems [11] Relationship between ridesharing network structure and benefits [12] Framework to model ridesharing and transit door-to-door services [14] Network traffic assignment model for ridesharing with real-time ridematching [15] Taxi models and architectures Centralized taxi scheduling model using real-time transfer system for ridesharing [16] Cloud-based architecture for taxi ridesharing system [17] Grid-based model and architecture for dynamic taxi ridesharing [18] Taxi scheduling model and architecture for dynamic ridesharing [19] Taxi-sharing problem as example of the Dial-a-Ride Problem (DARP) [20], [21], [22] Bus models and architectures Framework and architecture for largescale bus ridesharing/pooling [23] Dynamic bus routing model and simulator [24] Bus ridesharing model which combines different approaches (e.g., slugging and hitchhiking) [25] Autonomous vehicle models and architectures Space Time for Autonomous Ridesharing Systems (STARS) [26] P2P system model and shared autonomous fleet vehicles (SAFVs) for ridesharing [27] Ridesharing interactions between dynamic routing and optimization fleet size, road networks for shared autonomous vehicles [28] Discrete-time shared autonomous electric vehicles (SAEV) simulation model with rideshare matching [29] Universal Design (UD) model and shaping urban public transport with ridesharing [30] This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.…”

“…The authors in [28] identified strategic and operational challenges in shared autonomous vehicles which were addressed in several literature with issues relating to strategic problems treated exclusively for the operational challenges. The author proposed a unified optimization framework that considers foundationally the interactions between dynamic routing and optimization of variables such as fleet size, road networks and parking space allocation for shared autonomous vehicles.…”

Studying computer communication networks to identify errors using enhanced neural network algorithms

“…It should be noted that in reality, edge zero-filling may be applied to the original input matrix in order to guarantee smooth convolution window sliding. This process also aids in preventing dimensional mismatch when the convolution kernel's size exceeds the size of the input data 24 .…”

Economic and environmental benefits of automated electric vehicle ride-hailing services in New York City

“…With the use of simulation tools, studies have focused on congestion effects 7 , congestion pricing 8 , 9 , fleet re-balancing 10 , 11 , infrastructure planning 7 , 11 , 12 , environmental impact evaluation 13 , 14 . In the operations research approach, in addition to similar factors 15 , 16 , vehicle routing 17 , 18 , fleet sizing 19 , 20 , pickup and delivery 21 , 22 are considered. However, the two approaches exhibit distinct advantages and trade-offs.…”