Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Dynamic Strategic Plan for the Transition to a Clean Bus Fleet using Multi-Stage Stochastic Programming with a Case Study in Istanbul

Neman Karimi, Burak Kocuk, Tugce Yuksel

Abstract

In recent years, the transition to clean bus fleets has accelerated. Although this transition might bring environmental and economic benefits, it requires a long-term strategic plan due to the large investment costs involved. This paper proposes a multi-stage stochastic program to optimize strategic plans for the clean bus fleet transition that explicitly considers the uncertainty scenarios in the cost and efficiency improvements of clean buses. Our optimization model minimizes the total expected cost subject to emission targets, budget restrictions and several other operational considerations. We propose a new forecasting approach that captures the correlation between these improvements to obtain realistic future pathways for Battery Electric Buses (BEBs) and Hydrogen Fuel Cell Buses (HFCBs), which are then given to the multi-stage stochastic program as scenarios. We also utilize a physics-based model for BEBs to accurately capture their energy consumption and recharging needs. As a case study, we focus on the complex public bus network of Istanbul, which aims to transition to a clean bus fleet by 2050. Utilizing real datasets, we solve a five-stage stochastic program spanning a 25-year planning horizon that involves 256 scenarios to obtain dynamic strategic plans that can be used by the policy makers. Our results suggest that BEBs are more advantageous than HFCBs, even in slow BEB but fast HFCB development scenarios. We also conduct several sensitivity analyses to understand the effects of the intermediate emission targets, budget limitations and energy prices.

A Dynamic Strategic Plan for the Transition to a Clean Bus Fleet using Multi-Stage Stochastic Programming with a Case Study in Istanbul

Abstract

In recent years, the transition to clean bus fleets has accelerated. Although this transition might bring environmental and economic benefits, it requires a long-term strategic plan due to the large investment costs involved. This paper proposes a multi-stage stochastic program to optimize strategic plans for the clean bus fleet transition that explicitly considers the uncertainty scenarios in the cost and efficiency improvements of clean buses. Our optimization model minimizes the total expected cost subject to emission targets, budget restrictions and several other operational considerations. We propose a new forecasting approach that captures the correlation between these improvements to obtain realistic future pathways for Battery Electric Buses (BEBs) and Hydrogen Fuel Cell Buses (HFCBs), which are then given to the multi-stage stochastic program as scenarios. We also utilize a physics-based model for BEBs to accurately capture their energy consumption and recharging needs. As a case study, we focus on the complex public bus network of Istanbul, which aims to transition to a clean bus fleet by 2050. Utilizing real datasets, we solve a five-stage stochastic program spanning a 25-year planning horizon that involves 256 scenarios to obtain dynamic strategic plans that can be used by the policy makers. Our results suggest that BEBs are more advantageous than HFCBs, even in slow BEB but fast HFCB development scenarios. We also conduct several sensitivity analyses to understand the effects of the intermediate emission targets, budget limitations and energy prices.

Paper Structure

This paper contains 40 sections, 10 equations, 20 figures, 16 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Literature Review
  3. Our Approach and Contributions
  4. Methodology
  5. Problem Description
  6. Mathematical Formulation
  7. Energy Requirement Calculation
  8. Bus-to-Route Simulator
  9. Data Collection and Processing
  10. Public Bus Transit in Istanbul
  11. Initial Bus-Specific Costs
  12. Technological Change Forecasts
  13. BEB Technological Change
  14. HFCB Technological Change
  15. Scenario Tree Generation
  16. ...and 25 more sections

Figures (20)

  • Figure 1: BEB technology improvement charts. Cluster centers are marked with solid marks in (b) and (c), and can be distinguished by the abbreviations of Slow, Fast and Medium.
  • Figure 2: HFCB technology improvement charts.
  • Figure 3: Technology trees of DB ($B_1=1$), BEB ($B_2=2$) and HFCB ($B_3=2$) with $S=2$ stages.
  • Figure 4: Scenario tree with two stages.
  • Figure 5: Base Case results. Buses purchased are reported with bars and the cost figures are reported with markers.
  • ...and 15 more figures