201800010 5 ECTS Master of Science

Transportation and Logistics Management

Design, Scheduling, and Routing of Transportation Systems

Quantitative Approaches for Logistics Planning and Optimization

Breno Alves Beirigo, PhD

Course Coordinator

Industrial Engineering and Business Information Systems University of Twente Enschede, Netherlands

Date

2026-08-14

An introduction to transportation systems, optimization, and logistics decisions.

Canvas Syllabus Course Website

Course Overview

What You’ll Learn

This course examines the design, scheduling, and routing of different modes of transportation within supply chain management:

• Logistics and transportation management strategies
• Analysis and modeling of logistics systems
• Formulation of optimization problems
• Computer-aided planning methods (Python)
• Critical evaluation of solution approaches

Assessment

Component	Weight
Assignment 1	25%
Assignment 2	30%
Written Exam	45%

Pass requirements: Exam ≥5.5, Overall ≥5.5

Learning Objectives

By the end of this course, you will be able to:

1. Explain logistics and transportation management strategies
2. Analyze and describe logistics systems using analytical frameworks
3. Differentiate transportation problems within various scenarios
4. Formulate logistical issues from case studies
5. Design and apply solution approaches for logistics planning
6. Implement computer-aided planning methods in Python
7. Evaluate solution approaches critically
8. Communicate findings through structured reports

Course Structure

Weekly Activities

• 📚 8 Lectures
• 🎤 1 Guest Lecture
• 💻 4 Python Tutorials
• 🔧 8 Practical Sessions

Workload (140 hours)

• Contact hours: ~40h
• Assignments: ~50h
• Self-study: ~35h
• Exam prep: ~15h

Teaching Approach: Flipped classroom with pre-recorded videos, interactive sessions, and hands-on Python exercises.

Key Deadlines

Deadline	Date	Description
Assignment 1	March 7	Vehicle Routing Problem
Assignment 2	April 4	Network Design
Exam (1st)	April 15	3-hour written exam
Assignment Repairs	May 9	Resubmission deadline
Exam (2nd)	July 3	Retake opportunity

Important

Late submissions receive a 1-point penalty per day. Extensions only with valid documentation.

Why shortest paths matter in T&L

• Road freight routing: travel time / distance between stops (VRP subproblem)
• Urban logistics: bike couriers, one-way streets, turn restrictions
• Warehouse operations: AGV / picker routing on directed aisles
• Planning & control: repeated path queries → algorithm engineering matters

Shortest paths are a primitive used inside larger optimization models (VRP, min-cost flow, dynamic traffic assignment). Pallottino & Scutellà explicitly connect shortest path algorithms with transportation models. (Pallottino and Scutellà 1998)

The modeling object

A directed, weighted graph \(G=(V,E)\):

• nodes \(V\): depots, hubs, intersections, docks
• edges \(E\): roads/lanes/arcs (may be directed)
• cost \(w(u,v)\): distance, time, generalized cost (toll + time), etc.

Shortest-path query: find a minimum-cost path from origin \(s\) to destination \(t\).

Quote (problem framing): “We consider n points (nodes) … connected by a branch; the length of each branch is given.” (Dijkstra 1959)

Classics we’ll cover (and one important correction)

• Dijkstra: single-source shortest paths with nonnegative edge costs. (Dijkstra 1959)
• Bellman—Ford (context): handles negative edges (rare in physical travel time, but relevant in reduced costs / potentials). (Bellman 1958)
• Time-dependent shortest paths: edge costs depend on departure time (congestion, schedules). (Orda and Rom 1990; Dean 2004)
• Ford—Fulkerson: not a shortest-path algorithm — it solves max flow. It’s still a core network primitive in logistics. (Ford and Fulkerson 1956)

Users often bundle “graph algorithms” together. In logistics, shortest paths and max-flow/min-cut show up side-by-side (routing vs capacity/throughput).

Dijkstra: when and why

Requirements

• All edge weights \(w(e) \ge 0\)

Typical T&L uses

• shortest travel-time path in a city network
• repeated path calls in heuristics (e.g., insertion in VRP)
• as a subroutine in successive shortest path min-cost flow (after reweighting)

Complexity (common implementations)

• binary heap: \(O((|V|+|E|)\log |V|)\)
• Fibonacci heap: \(O(|E| + |V|\log |V|)\) (theoretical improvement) (Fredman and Tarjan 1987)

The Fibonacci-heap bound is a classic asymptotic result, but constant factors + engineering mean it’s not always faster in practice. (Fredman and Tarjan 1987)

A minimal “counterintuitive” directed streets example

One-way streets make “geographically closer” routes infeasible or expensive.

Figure 1: Directed graph showing one-way street constraints where the shortest path from S to T is not the most direct route.

• A tempting idea: go S→A→B→T (looks “direct”), but it costs 12
• The actual best path is S→C→T costing 4
• And A→C is not usable from C back to A (direction matters)

This slide is about directedness (one-ways, turn restrictions). In real road graphs, edges represent allowed movements; “reverse” may not exist or may have different cost.

Dijkstra step-by-step on the toy graph

We’ll compute shortest distances from S.

• Initialize: dist[S]=0, others ∞
• Repeatedly settle the smallest tentative distance, relax outgoing edges

Listing 1: Dijkstra’s algorithm implementation for the directed streets example, demonstrating the label-setting approach.

Step 1: settle S with dist=0
  relax S->A (w=1): dist[A]=1 prev[A]=S
  relax S->C (w=2): dist[C]=2 prev[C]=S
Step 2: settle A with dist=1
  relax A->B (w=1): dist[B]=2 prev[B]=A
Step 3: settle C with dist=2
  relax C->T (w=2): dist[T]=4 prev[T]=C
Step 4: settle B with dist=2
Step 5: settle T with dist=4

Final distances: {'S': 0, 'A': 1, 'B': 2, 'C': 2, 'T': 4}

Pedagogically: emphasize “settling” is valid because no negative edges exist; that’s the key correctness condition. (Dijkstra 1959; Cormen et al. 2009)

When Dijkstra is not enough

Negative edges (rare in literal travel time, common in transformed problems)

Examples:

• reduced costs after dual variables / potentials in min-cost flow
• “profit” arcs modeled as negative costs (careful: can create negative cycles)

If negative edges exist (but no negative cycles), use Bellman—Ford:

• Complexity: \(O(|V||E|)\) (Bellman 1958; Cormen et al. 2009)

In transportation, negative travel times are not physical, but algorithmic transformations can introduce negative reduced costs. Bellman’s routing problem is an early dynamic-programming view. (Bellman 1958)

Time-dependent shortest paths (TDSP)

Why TDSP appears in logistics

• peak/off-peak travel times
• time-of-day congestion
• schedules (e.g., ferries, shuttles), where waiting may be allowed

Model: each edge has a travel-time function \[ \tau_{uv}(t) = \text{time to traverse }(u,v)\text{ if departing at time }t \]

Key property: FIFO

• leaving later cannot arrive earlier (no “time travel”)
• many efficient TDSP algorithms assume FIFO (Dean 2004; Orda and Rom 1990)

Quote: “the delay (or weight) of the edges changes with time according to arbitrary functions…” (Orda and Rom 1990) Also: Dean’s FIFO framework is the standard modern reference for time-dependent Dijkstra-style methods. (Dean 2004)

A minimal TDSP example (piecewise congestion)

Nodes: \(S \rightarrow A \rightarrow T\) and direct \(S \rightarrow T\)

• \(S\rightarrow T\): constant 15 minutes
• \(S\rightarrow A\): constant 5 minutes
• \(A\rightarrow T\): depends on departure time \(t\)

\[ \tau_{AT}(t)= \begin{cases} 5, & t < 08{:}00\\ 20, & t \ge 08{:}00 \end{cases} \]

If you depart S at 07:55

• arrive A at 08:00 → then \(A\rightarrow T\) costs 20 → total 25 ⇒ choose direct (15)

If you depart S at 07:54

• arrive A at 07:59 → then \(A\rightarrow T\) costs 5 → total 10 ⇒ choose via A

This shows TDSP can make the “best path” departure-time dependent. A one-minute shift changes the optimal route. For FIFO networks, a Dijkstra-like label-setting approach can work with time-dependent arc evaluations. (Dean 2004)

Where Ford—Fulkerson fits (and why it’s here)

Ford—Fulkerson solves maximum flow: push as much as possible from source to sink given capacities. (Ford and Fulkerson 1956)

T&L examples:

• throughput of a warehouse conveyor network
• maximum number of trucks per hour through a constrained corridor (stylized)
• port/terminal capacity planning as a flow network

Relationship to shortest paths:

• different problem class (capacity feasibility vs cheapest route)
• but combined in min-cost flow and assignment problems (network flows toolbox) (Ahuja et al. 1993)

Important to be explicit: Ford—Fulkerson is not a shortest path algorithm; it’s a classic network optimization primitive in the same modeling family. (Ford and Fulkerson 1956; Ahuja et al. 1993)

Are Dijkstra / Ford—Fulkerson the “best and latest”?

They are foundational, not always best in practice for modern routing workloads.

What “best” depends on

• single query vs millions of queries
• graph type: arbitrary vs road-like (hierarchical, sparse, metric structure)
• static vs time-dependent vs frequently-updated costs

Speedups that dominate road routing

• A* (goal-directed search) (Hart et al. 1968)
• ALT (A* + landmarks + triangle inequality) (Goldberg and Harrelson 2005)
• Contraction Hierarchies (CH) (preprocess + tiny queries) (Geisberger et al. 2008)
• Customizable Contraction Hierarchies (CCH) for changing metrics (Dibbelt et al. 2016)
• Recent engineering + surveys: parallel CH construction and CCH refinements (Wan et al. 2024; Bläsius et al. 2025)

Quote (CH abstract): “We present a route planning technique solely based on the concept of node contraction.” (Geisberger et al. 2008) Quote (ALT abstract): “A* search … based on landmarks and the triangle inequality.” (Goldberg and Harrelson 2005)

“Reducing complexity” vs “showing speed gains”

Two different claims:

1. Asymptotic complexity improvements

• e.g., Fibonacci heaps improve Dijkstra’s bound on paper (Fredman and Tarjan 1987)

2. Orders-of-magnitude empirical speedups

• road networks exploit structure + preprocessing (ALT/CH/CCH/hub labels)
• typical result: far fewer settled nodes per query, sub-millisecond queries (after preprocessing) (Geisberger et al. 2008; Goldberg and Harrelson 2005; Bläsius et al. 2025)

If your workload is:

• few queries, small graphs → classic Dijkstra is often sufficient
• many queries, continent-scale → CH/CCH-style methods win decisively

Be careful in wording: some papers claim “faster” via experiments (engineering), not via better worst-case complexity. A good “least-reading” entry point is a survey, then pick one technique to implement. (Madkour et al. 2017; Bläsius et al. 2025)

Summary table

Table 1: Comparison of shortest-path algorithms and related network primitives in transportation and logistics.

Algorithm	Applies when	Time complexity (typical)	What it solves	Source
Dijkstra	edge weights nonnegative	\(O((|V|+|E|)\log |V|)\)	shortest paths	(Dijkstra 1959; Cormen et al. 2009)
Bellman—Ford	negative edges allowed (no neg cycles)	\(O(|V||E|)\)	shortest paths	(Bellman 1958; Cormen et al. 2009)
Time-dependent Dijkstra (FIFO)	FIFO time-dependent costs	depends on arc eval; label-setting works under FIFO	shortest paths by departure time	(Dean 2004)
Ford—Fulkerson	capacities on edges	depends on augmenting-path choices	max flow (not shortest path)	(Ford and Fulkerson 1956; Ahuja et al. 1993)
Contraction Hierarchies (speedup)	repeated queries on road-like graphs	preprocess + very fast queries	shortest paths (exact)	(Geisberger et al. 2008)

If you need one practical takeaway for logistics teaching:

• Dijkstra = baseline shortest path workhorse
• TDSP = realism for congestion/schedules
• CH/CCH = “how Google/OSRM-class routing gets fast”
• Ford—Fulkerson = capacity/throughput primitive that complements routing

References

Ahuja, Ravindra K., Thomas L. Magnanti, and James B. Orlin. 1993. Network Flows: Theory, Algorithms, and Applications. Prentice Hall.

Bellman, Richard. 1958. “On a Routing Problem.” Quarterly of Applied Mathematics 16 (1): 87–90.

Bläsius, Thomas et al. 2025. Customizable Contraction Hierarchies: A Survey.

Cormen, Thomas H., Charles E. Leiserson, Ronald L. Rivest, and Clifford Stein. 2009. Introduction to Algorithms. 3rd ed. MIT Press.

Dean, Brian C. 2004. Shortest Paths in FIFO Time-Dependent Networks: Theory and Algorithms. MIT Department of Computer Science.

Dibbelt, Julian, Ben Strasser, and Dorothea Wagner. 2016. “Customizable Contraction Hierarchies.” ACM Journal of Experimental Algorithmics 21 (1): 1.5:1–49. https://doi.org/10.1145/2886843.

Dijkstra, E. W. 1959. “A Note on Two Problems in Connexion with Graphs.” Numerische Mathematik 1 (1): 269–71. https://doi.org/10.1007/BF01386390.

Ford, L. R., and D. R. Fulkerson. 1956. “Maximal Flow Through a Network.” Canadian Journal of Mathematics 8: 399–404. https://doi.org/10.4153/CJM-1956-045-5.

Fredman, Michael L., and Robert E. Tarjan. 1987. “Fibonacci Heaps and Their Uses in Improved Network Optimization Algorithms.” Journal of the ACM 34 (3): 596–615. https://doi.org/10.1145/28869.28874.

Geisberger, Robert, Peter Sanders, Dominik Schultes, and Daniel Delling. 2008. “Contraction Hierarchies: Faster and Simpler Hierarchical Routing in Road Networks.” Proceedings of the 7th International Workshop on Experimental Algorithms (WEA), 319–33.

Goldberg, Andrew V., and Chris Harrelson. 2005. “Computing the Shortest Path: A* Search Meets Graph Theory.” Proceedings of the Sixteenth Annual ACM-SIAM Symposium on Discrete Algorithms (SODA), 156–65.

Hart, Peter E., Nils J. Nilsson, and Bertram Raphael. 1968. “A Formal Basis for the Heuristic Determination of Minimum Cost Paths.” IEEE Transactions on Systems Science and Cybernetics 4 (2): 100–107. https://doi.org/10.1109/TSSC.1968.300136.

Madkour, Amgad, Walid G. Aref, Faizan Ur Rehman, Mohamed Abdur Rahman, and Saleh Basalamah. 2017. A Survey of Shortest-Path Algorithms.

Orda, Ariel, and Raphael Rom. 1990. “Shortest-Path and Minimum-Delay Algorithms in Networks with Time-Dependent Edge-Length.” Journal of the ACM 37 (3): 607–25. https://doi.org/10.1145/79147.214078.

Pallottino, Stefano, and Maria Grazia Scutellà. 1998. “Shortest Path Algorithms in Transportation Models: Classical and Innovative Aspects.” In Equilibrium and Advanced Transportation Modelling, edited by Patrice Marcotte and Si Nguyen. Springer. https://doi.org/10.1007/978-1-4615-5757-9_11.

Wan, Z. et al. 2024. Parallel Contraction Hierarchies Can Be ...