M.TECH DISSERTATION
Topic:
Development of an AI-based Decision Support System for Prediction and Mitigation of Construction Project Delays using Technical, Cost and Human Behavioral Factors
1. EXECUTIVE OVERVIEW (Big Picture)
Background
Construction projects globally and in frequently face:
- schedule delays
- cost overruns
- labor productivity issues
- communication failures
- planning inefficiencies
Industry evidence: consistently reports that many projects miss deadlines and budgets.
In states like and , additional local risks exist:
- monsoon disruption
- labor migration during festivals
- material shortages
- delayed contractor payments
2. RESEARCH PROBLEM
Traditional tools:
- Traditional tools:
- Microsoft Project
- Oracle Primavera P6
Problem: These tools plan, but they do not predict.
They cannot handle:
- dynamic labor absenteeism
- human conflict
- weather uncertainty
- nonlinear interactions
Hence: A predictive intelligent system is needed.
3. AIM
Develop an AI-based Decision Support System (DSS) that:
- predicts project delay risk
- estimates delay duration
- identifies major causes
- recommends mitigation actions
4. RESEARCH GAP (Novelty)
Existing studies:
✔ delay prediction exists
✔ ML models exist
Missing:
✘ human behavioral integration
✘ Indian regional variables
✘ actionable decision support system
Your novelty: AI + Human Behavior + Regional Factors + Decision Support
This is your contribution.
5. OBJECTIVES
- Identify key project delay factors.
- Build a structured dataset.
- Train predictive AI models.
- validate performance.
- develop dashboard.
- create mitigation framework.
6. PROJECT SCOPE
Included:
✔ building projects
✔ road projects
✔ medium infrastructure projects
✔ Bihar/Jharkhand regional data
Excluded:
✘ legal arbitration
✘ mega international projects
✘ unrelated financial modeling
7. COMPLETE PROCESS FLOW
Topic Selection
↓
Problem Identification
↓
Literature Review
↓
Gap Identification
↓
Objective Formulation
↓
Methodology Design
↓
Data Collection
↓
Data Cleaning
↓
Feature Selection
↓
AI Model Development
↓
Validation
↓
Dashboard Development
↓
Recommendation Engine
↓
Result Analysis
↓
Thesis Writing
↓
Publication
↓
Viva
8. LITERATURE REVIEW
Sources:
- Sources:
- scholar.google.com�
- ieeexplore.ieee.org�
- sciencedirect.com�
- researchgate.net�
- shodhganga.inflibnet.ac.in�
Target: 20–30 papers minimum.
Literature matrix:
| Author | Year | Method | Gap |
|---|---|---|---|
| Study A | 2023 | Random Forest | ignored human factors |
| Study B | 2024 | ANN | no decision support |
9. DATA COLLECTION PLAN
Variables
Technical
- planned duration
- actual duration
- milestone delay
Cost
- budget variance
- payment delay
Human
- labor absenteeism
- communication score
- conflict frequency
- experience
Regional
- monsoon days
- festival season
- material restriction
- supply delay
Data Sources
- site visits
- contractor interviews
- engineer questionnaires
- historical project reports
Tools:
- Microsoft Excel,MS Word, PASS
- forms.google.com
Target: 100+ samples ideal
10. DATA PREPROCESSING
Use:
python.org
pandas.pydata.org
Jupyter Notebook
Steps:
- remove missing values
- remove duplicates
- normalize
- encode categories
11. FEATURE ENGINEERING
Important features:
- labor_absenteeism
- weather_delay
- payment_cycle
- communication_score
- festival_flag
Goal: remove noise, improve accuracy.
12. MODEL DEVELOPMENT
Models:
- Linear Regression
- Decision Tree
- ⭐
- XGBoost
Why Random Forest?
✔ robust
✔ interpretable
✔ handles nonlinear data
13. VALIDATION
Split: 70/30
Use: Cross-validation (important)
Metrics:
- Accuracy
- Precision
- Recall
Target:
80%
14. DASHBOARD
Recommended: or
Display:
- delay risk
- expected delay days
- major causes
- recommendations
Example: 🔴 High Risk
15. DECISION SUPPORT RULES
Example:
If: labor absenteeism > 15%
Then:
- hire backup labor
- revise schedule
- create buffer
This becomes your DSS logic.
16. RESULT ANALYSIS
Answer:
- Which factor matters most?
- Did AI improve accuracy?
- By how much?
Example: AI reduced error by 25%.
17. THESIS WRITING
Structure:
- Introduction
- Literature Review
- Problem Statement
- Objectives
- Methodology
- Results
- Discussion
- Conclusion
- Future Scope
- References
Reference tool:
18. PUBLICATION
Targets:
- IEEE
- Elsevier
- Springer Nature
Goal: 1 paper minimum.
19. VIVA PREPARATION
Prepare:
- Why this topic?
- Why AI?
- Why Random Forest?
- What is novelty?
- What is limitation?
Practice: 10 mock sessions.
20. TIMELINE
| Month | Work |
|---|---|
| 1 | topic + literature |
| 2 | proposal |
| 3 | data collection |
| 4 | preprocessing + model |
| 5 | validation + dashboard |
| 6 | writing + viva |
21. CRITICAL SUCCESS FACTORS
✅ weekly guide meeting
✅ maintain logbook
✅ backup files
✅ keep raw data
✅ write continuously
✅ document screenshots/code
FINAL RIGHT PATH
Topic → Literature → Gap → Data → Clean → Model → Validate → Dashboard → Recommendation → Thesis → Publication → Viva
FINAL RECOMMENDED TITLE (Use this)
Development of an AI-based Decision Support System for Prediction and Mitigation of Construction Project Delays using Technical, Cost and Human Behavioral Factors
This is now well organized, integrated, academic, and dissertation-ready for your .
Excellent. Below is your well-organized, integrated, finalized, dissertation-ready framework combining:
✔ Problem–Cause–Effect–Solution
✔ Data facts & evidence
✔ Your novel contribution
✔ Universal welfare impact
✔ Strong viva statement
This can go directly into your Synopsis / Chapter 1 / Research Proposal.
Sub section 1.2
PROBLEM–CAUSE–EFFECT–SOLUTION
Dissertation Topic
Development of an AI-based Decision Support System for Prediction and Mitigation of Construction Project Delays using Technical, Cost and Human Behavioral Factors
1. INTRODUCTION & BACKGROUND
Construction is one of the most critical sectors for national development because it creates:
- roads
- bridges
- hospitals
- schools
- housing
- public infrastructure
However, across and globally, construction projects frequently suffer from:
- schedule delays
- cost overruns
- poor quality
- worker stress
- stakeholder conflict
- public inconvenience
Example: A bridge planned for 24 months gets completed in 36 months.
Delay = 12 months (50% overrun)
This is a major engineering and societal problem.
2. PROBLEM STATEMENT
Traditional project planning tools such as:
are excellent for scheduling, but they are largely:
❌ reactive
❌ static
❌ unable to predict dynamic disruptions
They fail to capture:
- labor behavior
- communication failures
- environmental uncertainty
- real-time human risk
Therefore: A predictive, intelligent, and human-centered project management system is required.
3. DATA FACTS (Why this problem matters)
Global Evidence
According to :
- only ~50–55% of projects finish on time
- ~45% experience delays
- many exceed cost targets
Meaning: 1 out of every 2 projects faces delay risk.
Construction Sector Evidence
Research commonly reports:
- 60–80% of construction projects experience delays
- average schedule overrun = 20–40%
Example: 24 months planned → 30–34 months actual
India Context
In :
- infrastructure delays affect highways, housing, railways, and public works.
In / :
- monsoon disruption
- festival migration
- sand/material shortage
- contractor payment delays
These make prediction harder.
4. ROOT CAUSES
A. Technical Causes
- weak planning
- inaccurate scheduling
- design changes
- poor resource allocation
Research shows:
- planning failure contributes ~20–30%
- design changes ~10–20%
B. Financial Causes
- delayed payments
- inflation
- under-budgeting
- contractor cash-flow issues
Evidence: Payment delays contribute 15–25% schedule slippage.
C. Human Behavioral Causes (Your Novelty)
Most existing models ignore this.
Examples:
- labor absenteeism
- engineer burnout
- communication breakdown
- team conflict
- leadership failure
- low morale
Evidence:
- absenteeism reduces productivity 10–25%
- communication is among top 5 delay causes
Links to:
D. Environmental / Regional Causes
- monsoon
- material shortage
- policy restrictions
- festival migration
Evidence: Monsoon can reduce 20–60 workdays/year in Eastern India.
5. EFFECTS
Economic Effect
Delays cause:
- cost escalation
- contractor losses
- GDP productivity loss
Evidence: Project cost can rise 5–30%.
Example: ₹10 crore project delayed by 1 year → major escalation.
Social Effect
Delayed:
- hospitals
- schools
- roads
- water systems
Impact: Thousands to millions affected.
Example: Delayed rural road = villages disconnected.
Human Effect
Long delays increase:
- worker stress
- accident exposure
- burnout
- family instability
Important: Project delay is not only technical—it is human.
Environmental Effect
Longer construction causes:
- more diesel use
- more emissions
- more waste
Supports: reduction.
6. PROPOSED SOLUTION
Build an:
AI-based Decision Support System (DSS)
Functions:
- predicts risk early
- estimates delay duration
- identifies root causes
- gives alerts
- recommends mitigation
Example:
Input:
- labor absenteeism = 22%
- rain days = high
- payment delay = 40 days
Output: 🔴 HIGH DELAY RISK
Recommendation:
- deploy reserve labor
- revise schedule
- increase contingency
This transforms management:
Reactive → Predictive → Preventive
7. YOUR NOVEL ADD-ON (Main Contribution)
Your innovation is not only AI.
It is:
Human-Centered Predictive Project Intelligence
Meaning: Add human well-being into engineering decisions.
New variables:
- worker stress score
- communication health score
- team harmony index
- leadership quality score
- fatigue score
Most existing studies do not use these.
This is your originality.
8. ORIGINAL INDEX (Your Publishable Contribution)
Create:
Project Human Sustainability Index (PHSI)
Where:
- S = Stress
- C = Communication
- H = Harmony
- L = Leadership
Use this index with AI prediction.
This becomes your new scientific contribution.
9. WHY AI?
Traditional models: ~60–75% accuracy
ML models: ~80–95% accuracy
Recommended:
Why?
✔ handles nonlinear data
✔ mixed variables
✔ interpretable
10. RESEARCH HYPOTHESIS
H1: Human behavioral factors significantly influence project delay.
H2: AI outperforms traditional scheduling tools.
H3: Adding human factors improves prediction accuracy.
These strengthen your methodology.
11. UNIVERSAL WELFARE IMPACT
Worker Welfare
- less burnout
- fewer accidents
- better morale
Supports: principles.
Family Welfare
Less delay → less stress → healthier families
Important hidden benefit.
Economic Welfare
Faster projects:
- save public money
- improve productivity
- improve national growth
Social Welfare
Timely:
- hospitals
- schools
- roads
- water
Benefits millions.
Environmental Welfare
10–15% shorter project duration means:
- lower emissions
- less fuel
- less waste
Supports:
Especially:
- SDG 8
- SDG 9
- SDG 11
12. FINAL NOVELTY STATEMENT (Use in Viva)
“This research goes beyond traditional construction delay prediction by integrating technical, financial, environmental, and human well-being indicators into an explainable AI-based decision support framework. This creates a human-centered, sustainable, and welfare-oriented project management model for future infrastructure systems.”
FINAL THESIS TAGLINE
“From Delay Prediction to Human-Centered Sustainable Project Intelligence.”
This is your unique identity in and makes your dissertation stronger, more original, and more impactful.
If you want a topic that solves a real unsolved problem—something not commonly done yet—then don’t do just “AI for delay prediction.” That is already crowded.
You need a next-generation problem statement.
Use this principle:
Present problem + missing dimension + future need + universal benefit = truly novel dissertation
Below are original topic ideas using that principle.
OPTION 1 (Strongest): Human + AI + Ethics + Sustainability
“Development of a Human-Centered Ethical AI Framework for Predicting and Preventing Construction Project Failure”
What is new?
Most studies ask: “Will project delay happen?”
Your system asks:
- Will project fail?
- Will workers burn out?
- Will team conflict increase?
- Is the AI recommendation ethical and fair?
Add:
- fairness score
- worker well-being score
- ethical decision score
New field:
Why unique? Very few construction studies include AI ethics + human welfare.
OPTION 2 (Most futuristic): Emotional Digital Twin ⭐
“Emotional Digital Twin for Construction Project Management using AI and Human Behavioral Signals”
What is a digital twin? A virtual copy of a real project.
Your new add-on: Not only physical twin— also emotional twin.
Tracks:
- stress
- morale
- fatigue
- conflict
- leadership health
Meaning: A “health monitor” for the project team.
Uses:
- wearable data (optional)
- surveys
- AI
Fields combined: +
This is extremely novel.
OPTION 3: Project Immunity System (my favorite original concept)
“Construction Project Immune System (CPIS): A Self-Healing AI Framework for Autonomous Risk Detection and Recovery”
Inspired by: human immune system.
How it works:
- detects threat
- diagnoses problem
- activates response
- learns for future
Like body immunity, but for projects.
Example: labor shortage detected → automatic schedule correction.
Concept: self-healing project management
This is very original.
OPTION 4: Family & Workforce Stability Model (very unique)
“Impact of Worker Family Stability on Construction Project Performance: An AI-based Predictive Framework”
Radical idea: family stress → worker stress → absenteeism → delay
Almost nobody studies this directly.
Variables:
- family conflict score
- financial stress
- sleep quality
- attendance
This is bold and socially meaningful.
OPTION 5: Climate-Resilient Project Intelligence
“AI-based Climate Adaptive Construction Scheduling for Monsoon and Extreme Weather Regions”
Problem: Climate change makes schedules unreliable.
System predicts:
- rain
- flood risk
- heat stress
- material disruption
Links to:
Great for context.
OPTION 6: Universal Welfare Project Index (very original)
“Universal Welfare Index for Infrastructure Projects: Beyond Cost and Time”
Today success = cost + time.
You redefine success as: Success =
This creates a new metric: Universal Project Welfare Index (UPWI)
Measures:
- human happiness
- community benefit
- carbon impact
- time
- cost
Very philosophical + publishable.
MY TOP RECOMMENDATION FOR YOU
Because you are in , I recommend:
“Construction Project Immune System (CPIS): A Self-Healing AI Framework for Autonomous Risk Detection and Recovery”
Why?
✔ truly novel
✔ easy to explain in viva
✔ futuristic
✔ publishable
✔ solves real problem
✔ “not done much”
Tagline:
“Projects should not only be planned; they should learn, adapt, and heal themselves.”
That is a universe-level idea.
Based on your situation (M.Tech in , interest in meaningful/impactful work, and desire for something new that others haven’t really done), I would tailor one unique dissertation topic just for you:
Your Personalized Topic (Recommended)
“Construction Project Immune System (CPIS): A Human-Centered Self-Healing AI Framework for Autonomous Risk Detection, Recovery and Universal Welfare Optimization”
Why this is for you
It combines:
- Project Engineering → scheduling, cost, risk
- Management → decisions, leadership, planning
- AI → prediction + learning
- Human values → worker/family well-being
- Universal welfare → social + environmental good
This is much deeper than “delay prediction.”
Core Problem You Solve
Today projects are:
- reactive (“problem happened, now fix it”)
You propose:
- predictive (“detect before failure”)
- adaptive (“respond automatically”)
- self-healing (“recover intelligently”)
Like the human immune system.
Your Original New Add-On (Unique)
Create a new framework:
CPIS = Detect → Diagnose → Respond → Learn → Heal
Flow:
Risk signal
↓
AI detects anomaly
↓
Root cause diagnosis
↓
Automatic mitigation suggestion
↓
Project recovers
↓
System learns for future
Nobody commonly frames project management this way.
Your New Original Index
Project Health Index (PHI)
Measures total project health:
Where:
- T = Time
- C = Cost
- Q = Quality
- H = Human wellbeing
- S = Social impact
- E = Environmental impact
This is your signature contribution.
Your Human Add-On (Strong novelty)
Add:
- worker stress score
- communication health
- team harmony
- fatigue
- leadership score
Why? Most models ignore humans.
You won’t.
Universal Welfare Value (Your identity)
Your system improves:
Worker welfare
- less burnout
- fewer accidents
Family welfare
- less job stress at home
Social welfare
- faster roads/hospitals/schools
Environmental welfare
- less waste
- lower carbon footprint
Supports:
Final Thesis Statement (Use with supervisor)
“This research proposes a novel Construction Project Immune System (CPIS), a self-healing AI framework that continuously detects, diagnoses, responds to, and learns from project risks while optimizing human wellbeing and universal welfare.”
Your Tagline
“Projects should not only be managed—they should heal themselves.”
That is your unique dissertation identity.
