M.TECH (PROJECT ENGINEERING & MANAGEMENT

 

M.TECH (PROJECT ENGINEERING & MANAGEMENT)

Final Integrated Research

Enhanced, Publication-Oriented & PhD-Ready Version

DISSERTATION TITLE

Development and Validation of an AI-Enabled Integrated Project

Performance Optimization Framework Using Risk Management,

Resource Optimization, Earned Value Management and

Project Success Index for Engineering Projects

Prepared by:

Vimal Noble

Faculty & M.Tech Scholar, BIT Sindri

Jharkhand University of Technology (JUT) Programme:

M.Tech PEM — Batch 2024–26

Institute:

JUT Ranchi harkhand

 1. Research Background

Engineering projects across construction, manufacturing, infrastructure, energy, transportation, and industrial sectors continue to experience persistent delivery challenges. Global studies from organisations such as the Project Management Institute (PMI) and Independent Project Analysis (IPA) confirm that a significant proportion of engineering projects fail to achieve planned cost, time, and quality objectives.

Key Execution Challenges

• Cost overruns and budget escalation

• Schedule delays and milestone slippage

• Resource inefficiencies and allocation failures

• Quality defects, rework, and productivity losses

• Safety incidents and poor decision-making

The Silo Problem

Current project management approaches treat critical performance dimensions as separate, isolated functions:

Dimension Current Approach Limitation

Risk Management Standalone risk registers No integration with cost/schedule

Resource Management Manual allocation sheets No predictive optimization

Schedule Management CPM/Gantt charts only Lagging, not predictive

Cost Control Periodic budget reviews Reactive, not prescriptive

Quality Management Defect tracking in silos Disconnected from project health

This research proposes a unified framework integrating all critical project dimensions through AI-assisted decision support and a novel Project Success Index (PSI).

2. Problem Statement

Technical Problems

• Incomplete risk assessment models lacking probabilistic accuracy

• Inefficient resource allocation leading to idle time and resource leveling constraints

• Lagging schedule monitoring indicators that capture past errors rather than future risks

• Compounding cost escalation and expensive structural rework / quality failures

Management Problems

• Fragmented decision support systems that cannot process interdependencies

• Lack of predictive and prescriptive analytics at the execution level

• Delayed corrective actions due to slow dashboard updates

• Absence of an overarching, integrated project health performance indicator

Research Gap

Existing literature primarily operates in silos, focusing on individual domains:

  • Risk Management only

  • Earned Value Management (EVM) only

  • Resource Optimization only

  • Isolated AI Applications

Very few studies integrate ALL project performance dimensions into a single, cohesive, statistically validated framework.

3. Research Aim & Objectives

Research Aim

To develop and validate an AI-enabled integrated project performance optimization framework capable of improving engineering project success.

Research Objectives

1. Identify critical risk factors across diverse engineering project lifecycles.

2. Prioritize and rank identified risks using Fuzzy Analytical Hierarchy Process (Fuzzy AHP).

3. Evaluate resource utilization efficiency using optimization heuristics.

4. Assess integrated schedule, cost, and quality performance parameters simultaneously.

5. Develop an AI-based predictive decision-support layer (LSTM / XGBoost).

6. Formulate a unified Project Success Index (PSI) including a structural risk penalty function.

7. Validate the comprehensive framework using Structural Equation Modeling (SEM).

8. Propose actionable implementation guidelines for industrial project deployment.

4. Research Questions & Hypotheses

Research Questions

RQ# Research Question

RQ1 How does proactive Risk Management influence final Project Success?

RQ2 How does Resource Optimization affect field-level Project Performance?

RQ3 What is the mathematical relationship between EVM metrics (SPI, CPI) and overall Project Success?

RQ4 Can AI-based predictive decision support significantly mitigate variance and improve project outcomes?

RQ5 Can a composite Project Success Index (PSI) provide a reliable and comprehensive measure of project health?

RQ6 What is the combined, systemic effect of RM, RO, SPI, CPI and QM on project success?

Research Hypotheses

Hypothesis Type Statement

H1 Direct Effect Risk Management positively influences Project Success.

H2 Direct Effect Resource Optimization positively influences Project Success.

H3 Direct Effect Schedule Performance positively influences Project Success.

H4 Direct Effect Cost Performance positively influences Project Success.

H5 Direct Effect Quality Management positively influences Project Success.

H6 Moderating Effect AI-Based Decision Support positively moderates the relationship between project management practices and project success.

5. Integrated Conceptual Framework

The framework integrates five project management dimensions, mediated by an AI Decision Support System, producing a composite Project Success Index (PSI).

[ PROJECT INPUTS ]

▼

FIVE INTEGRATED PROJECT MANAGEMENT DIMENSIONS

Risk Management (RM) • Resource Optimization (RO) • Schedule Management (SM)

Cost Management (CM) • Quality Management (QM)

▼

AI DECISION SUPPORT SYSTEM (AI-DSS)

Forecasting: LSTM Networks • Classification: XGBoost

Optimization: Genetic Algorithm (GA) • Particle Swarm Optimization (PSO)

▼ (Moderating Effect)

PROJECT SUCCESS INDEX (PSI)

Composite Health Metric & Performance Classification Thresholds

▼

[ SUSTAINED PROJECT SUCCESS ]

6. Research Variables

Independent Variables (IV)

• Risk Management (RM): Risk Identification accuracy, assessment consistency, risk response velocity

• Resource Optimization (RO): Labour efficiency, equipment utilization rates, material waste minimization

• Schedule Management (SM): Planning accuracy, schedule variance tracking, delay path mitigation

• Cost Management (CM): Budget compliance, cost tracking automation, cash-flow variance control

• Quality Management (QM): Non-conformance/rework rate, defect density, technical specification adherence

Moderator Variable

• AI-Based Decision Support (AI): Predictive scheduling, proactive risk alerts, resource and cost forecasting accuracy

Dependent Variable (DV)

• Project Success (PS): Measured empirically through cost performance, schedule milestones, quality indices, and multi-stakeholder satisfaction

7. Core Engineering & Project Management Formulas

Risk Priority & Evaluation

Risk Score RS = P × I

P = Probability of Occurrence | I = Impact Severity

Risk Priority Number RPN = S × O × D

S = Severity | O = Occurrence | D = Detection

Earned Value Management (EVM)

Schedule Performance Index SPI = EV / PV

EV = Earned Value | PV = Planned Value

Cost Performance Index CPI = EV / AC

EV = Earned Value | AC = Actual Cost

Operational Efficiencies

Resource Utilization RU (%) = (Actual Hours / Available Hours) × 100

Productivity Index Productivity = Output / Input

Quality Index QI (%) = (Accepted Work / Total Work) × 100

8. AI-Based Decision Support & Optimization Layer

Predictive Analytics Layer

Model Type Primary Function Key Output Metrics

LSTM Networks Deep Learning (Time-Series) Forecast schedule delays and cost overruns from historical sequences RMSE, MAE, Forecast Accuracy %

XGBoost Gradient Boosting (Ensemble) Predict risk event occurrence and classify project performance trajectories AUC-ROC, F1-Score, Feature Importance

Optimization Algorithms

Algorithm Type Optimization Target Application Domain

Genetic Algorithm (GA) Evolutionary Metaheuristic Resource allocation, project scheduling, critical path optimization Labour, equipment, material leveling

Particle Swarm (PSO) Swarm Intelligence Equipment utilization, multi-resource leveling, workforce deployment Fleet, workforce, cost optimization

AI-DSS Data Flow

Historical & Live Project Data

         ↓

LSTM Forecasting ←→ XGBoost Risk Classification

         ↓

GA / PSO Optimization Engines

         ↓

Optimized Project Plan → PSI Score → Decision Alert

9. Novel Academic Contribution: Project Success Index (PSI)

This research presents a comprehensive, mathematically balanced index to replace one-dimensional project health indicators. The PSI consolidates all critical performance dimensions into a single composite score.

PSI Formula

PSI = β₁(RM) + β₂(RO) + β₃(SPI) + β₄(CPI) + β₅(QM) − Φ(Risk Penalty)

β values derived from SEM path coefficients combined with Fuzzy AHP priority vectors

Φ = Non-linear Risk Penalty Function, activated when RS ≥ critical threshold

PSI Performance Classification Matrix

PSI Range Classification Action Required

90 – 100 Excellent Maintain current execution path; document as benchmark practices.

80 – 89 Very Good Normal monitoring; optimize remaining micro-level resource lags.

70 – 79 Good Minor corrective actions needed for lagging indicators.

60 – 69 Average Focused intervention required; high potential for variance leakage.

< 60 POOR — Critical Distress Structural framework intervention mandatory. Escalate immediately.

10. Research Methodology

Phase I — Literature Review & Synthesis

• Systematic review across IEEE Xplore, Elsevier ScienceDirect, Springer Nature, Taylor & Francis, Wiley, Scopus, and Google Scholar

• Establishment of the fundamental construct pool for all research variables

Phase II — Questionnaire & Instrument Design

• Development of Likert-scale psychometric and quantitative questionnaire

• Target Profile: Project Directors, Project Managers, Planning Engineers, Consultants, and Contractors across major infrastructure / engineering sectors

Phase III — Data Collection & Sampling Strategy

Sample Size Benchmark Status Purpose

150–200 Respondents Acceptable Baseline Structural validity confirmed

250+ Respondents Preferred Threshold Reliable SEM estimation

300+ Respondents Ideal (Excellent) Deep multi-group analysis (MGA)

Sampling Strategy: Combined Purposive Sampling and Snowball Sampling to target domain experts across engineering project sectors.

Phase IV — Statistical Analytics & Software Toolkit

Analytical Domain Tool / Software Core Metrics

Risk Matrix & Prioritization Expert Choice, Python, Excel Fuzzy AHP consistency index, FMEA parameters

Scheduling & Resource Baselines Primavera P6, MS Project Critical Path Method (CPM), Resource Histograms

Measurement Model Validation SmartPLS, AMOS Cronbach's α (≥0.7), AVE (≥0.5), CR (≥0.7), HTMT (<0.85)

Structural Modeling (SEM) SmartPLS, AMOS Path Coefficients (β), R², f² effect size, Q² predictive relevance

AI Modeling & Optimization Python (Scikit-Learn, TensorFlow) RMSE, MAE, GA Fitness Functions, PSO convergence vectors

11. Staged Execution Plan (Semester-Wise)

Semester Focus Area Key Deliverables Publication Target

Semester I Core Literature, Variable Freezing, Exploratory Risk Modeling Risk Registry, Fuzzy AHP Model, Mini Project Report Conference Paper 1

Semester II Questionnaire Maturation, Expert Validation, Pilot Study Finalized Survey Instrument, P6 Baseline Schedules Instrument Pilot Report

Semester III Large-Scale Data Collection, SEM Analysis, AI Model Development SEM Path Models, LSTM/XGBoost Architectures, GA Routines Conference Paper 2 + Journal Paper 1

Semester IV Case Study Implementation, PSI Deployment, Thesis Defense Prep M.Tech Master Dissertation Document Journal Paper 2 + Thesis Paper

Mini Project — Semester I

Title: Risk Identification and Prioritization in Engineering Projects Using Fuzzy AHP

Objectives: Identify major project risks • Rank risks using FAHP • Develop risk hierarchy • Validate expert opinions

Deliverables: Comprehensive Risk Registry, validated Fuzzy AHP Model, Mini Project Report

Publication: Conference Paper 1 — Application of Fuzzy AHP for Risk Prioritization in Engineering Projects

12. Publication & Deliverables Roadmap

Research Output Inventory

Research Artifact Quantity Semester Academic Value / Focus

Master Dissertation Thesis 1 IV Comprehensive integration of all components and findings

Mini Project Report 1 I Foundation risk baseline via Fuzzy AHP

Conference Paper 1 1 I Risk Prioritization using Fuzzy AHP in Engineering Projects

Conference Paper 2 1 III Resource Optimization using Genetic Algorithms (GA/PSO)

Journal Paper 1 (Scopus/SCI) 1 III AI-Enabled Project Performance — LSTM/XGBoost Architecture

Journal Paper 2 (Scopus/SCI) 1 IV Development and Validation of Project Success Index (PSI)

Fuzzy AHP Risk Model 1 I Computational risk prioritization engine

GA Resource Optimization Model 1 III Metaheuristic resource leveling solution

LSTM Variance Forecaster 1 III Time-series cost/schedule forecasting model

Project Success Index (PSI) 1 IV Composite project health index framework

Strategic Assessment

Overall Research Architecture Quality: PhD-Ready Level

This framework pairs robust multivariate structural equations (SEM) with advanced machine learning (LSTM / XGBoost / GA / PSO), moving beyond standard project monitoring into predictive and prescriptive domain optimization.

Final Publication Pathway:

  2 Conference Papers + 2 Scopus/SCI Journal Papers + 1 Integrated Thesis + 3 Functional Computational Models

This forms an ideal foundation for a future doctoral (PhD) research transition.