Introduction to Operational Research: Definition, Methods & Applications

TAKEAWAYS

• Operational Research (OR) is a scientific approach to decision-making using mathematical models to optimize complex systems.
• The general OR methodology follows a structured cycle: problem formulation → model building → solution → implementation → review.
• OR uses tools like linear programming, queuing theory, and game theory to solve real-world problems efficiently.
• Advantages include cost savings, improved efficiency, and data-driven decision-making, but limitations include high computational costs and model inaccuracies.
• Nepalese examples like NTC’s network optimization and Daraz’s inventory management demonstrate OR’s practical use.
• The OR process is iterative, requiring continuous feedback for refinement.

1. Definition and Scope of Operational Research

Operational Research (OR) is a discipline that applies scientific methods, mathematical modeling, and algorithms to solve complex, real-world problems in business, industry, government, and everyday life. It bridges the gap between theory and practice by transforming ill-structured problems into structured, solvable models.

Key Characteristics of OR

mindmap
  root((Operational Research))
    - Applied Science
      - Uses math, statistics, and logic
      - Focuses on real-world problems
    - Interdisciplinary
      - Combines engineering, economics, and management
    - Problem-Solving Approach
      - Structured methodology
      - Iterative process
    - Decision-Making Tool
      - Optimizes resources
      - Reduces costs and delays

Difference Between OR and Traditional Problem-Solving

2. The General Methodology of Operational Research

The OR process follows a cyclical approach to ensure continuous improvement. The steps are:

1. Problem Identification

   • Define the problem clearly (e.g., "How to reduce Daraz’s delivery delays?").
   • Consult stakeholders (e.g., logistics managers, customers).

2. Model Formulation

   • Simplify the problem into a mathematical or logical model.
   • Example: For Daraz, model delivery routes as a network flow problem.

3. Solution

   • Use OR tools (e.g., linear programming, simulation) to find optimal solutions.
   • Example: Solve for the shortest delivery path using the shortest-path algorithm.

4. Implementation

   • Deploy the solution in the real world (e.g., adjust Daraz’s delivery schedules).

5. Review and Feedback

   • Monitor results and refine the model if needed (e.g., adjust for traffic patterns).

flowchart TD
    A["Problem Identification"] --> B["Model Formulation"]
    B --> C["Solution"]
    C --> D["Implementation"]
    D --> E["Review & Feedback"]
    E --> A

Worked Example: Optimizing NTC’s Network Routing

Problem: NTC wants to reduce call drop rates by optimizing its network routes. OR Approach:

1. Model: Represent the network as a graph where nodes = cell towers, edges = signal paths.
2. Objective: Minimize signal loss by optimizing tower placement.
3. Solution: Use graph theory to find the shortest path for signal transmission.
4. Implementation: Deploy new towers along optimal routes.
5. Review: Monitor call quality improvements.

3. Tools and Techniques of Operational Research

OR uses a variety of mathematical and computational tools to solve problems. Key techniques include:

A. Mathematical Models

• Linear Programming (LPP): Optimizes linear objectives (e.g., maximizing profit, minimizing cost).
• Network Models: Used for routing, scheduling (e.g., Pathao’s driver assignment).
• Inventory Models: Manages stock levels (e.g., Daraz’s warehouse optimization).

B. Statistical Techniques

• Regression Analysis: Predicts trends (e.g., Ncell’s customer churn prediction).
• Simulation: Models complex systems (e.g., Kathmandu traffic flow simulation).

C. Heuristics and Metaheuristics

• Approximate solutions for large problems (e.g., genetic algorithms for NEPSE stock portfolio optimization).

D. Game Theory

• Analyzes strategic interactions (e.g., NTC vs. Ncell’s network pricing wars).

E. Queuing Theory

• Models waiting lines (e.g., Khalti’s transaction processing queue).

4. Advantages and Limitations of OR

Advantages

• Cost-Effective: Reduces waste (e.g., banks saving on loan processing).
• Data-Driven: Uses real data for decisions (e.g., eSewa’s fraud detection).
• Scalable: Solutions adapt to growth (e.g., Daraz’s inventory scaling).
• Improves Efficiency: Optimizes processes (e.g., NTC’s network uptime).

Limitations

• High Computational Cost: Some models require powerful software.
• Model Assumptions: Real-world problems are rarely perfectly modeled.
• Implementation Challenges: Resistance to change in organizations.
• Ethical Concerns: OR can be misused for unethical decisions (e.g., price discrimination).

5. Real-World Applications in Nepal and Globally

In the Real World

1. eSewa’s Fraud Detection

   • Idea: Uses statistical anomaly detection to flag suspicious transactions.
   • How: OR models analyze transaction patterns to identify fraud in real time.

2. Daraz’s Inventory Management

   • Idea: Inventory control models (e.g., EOQ) optimize stock levels.
   • How: Balances holding costs vs. stockout risks to reduce losses.

3. Pathao’s Driver Assignment

   • Idea: Assignment problem matches drivers to optimal routes.
   • How: Minimizes idle time and fuel costs using linear programming.

4. NEPSE’s Stock Portfolio Optimization

   • Idea: Multi-objective optimization balances risk and return.
   • How: OR models help investors diversify portfolios efficiently.

5. NTC’s Network Optimization

   • Idea: Graph theory optimizes signal routing.
   • How: Reduces call drops by strategically placing cell towers.

6. Khalti’s Transaction Processing

   • Idea: Queuing theory manages transaction queues.
   • How: Predicts wait times to improve user experience.

6. Limitations of Operational Research

While OR is powerful, it has practical constraints:

• Assumptions vs. Reality: Models often simplify complex systems.
• Data Quality: Garbage in, garbage out (e.g., inaccurate Daraz sales data).
• Human Factors: Employees may resist automated solutions.
• Ethical Issues: OR can be used for unethical pricing (e.g., dynamic pricing in airlines).

7. Worked Example: Loan Interest Calculation for a Bank

Problem: A bank wants to maximize profit from loans while keeping interest rates competitive.

OR Approach

1. Model:

   • Let $x$ = loan amount, $r$ = interest rate.
   • Objective: Maximize profit $P=x\cdot r\cdot t$ (where $t$ = time).
   • Constraints:
     • $r\le r_{\text{max}}$ (regulatory limit).
     • $x\le x_{\text{max}}$ (bank’s lending capacity).

2. Solution:

   • Use linear programming to find optimal $x$ and $r$.
   • Example: If $r_{\text{max}}=10\%$ and $x_{\text{max}}=\$1M$, solve: $$\text{Maximize }P=1,000,000\times r\times 1\text{ year}$$ Subject to $r\le 0.10$.

3. Result:

   • Optimal $r=10\%$, $P=\$100,000$.

8. Comparison Table: OR vs. Traditional Management

Exam Tip

• For short-answer questions (2.5+2.5 marks):

  • Define OR in 1-2 sentences (e.g., "OR is the application of scientific methods to solve complex problems").
  • Mention 2-3 key tools (e.g., linear programming, queuing theory).
  • Briefly explain the OR methodology (problem → model → solve → implement → review).
  • Highlight 1 advantage and 1 limitation (e.g., "OR reduces costs but requires high computational power").

• For long-answer questions (5+5 marks):

  • Start with a clear definition of OR.
  • Explain the step-by-step OR methodology with a real-world example (e.g., Daraz’s inventory).
  • Discuss 2 tools/techniques and their applications.
  • End with limitations and why OR is still valuable.

• Common Pitfalls:

  • Overcomplicating definitions (stick to concise explanations).
  • Ignoring real-world examples (always tie theory to Nepalese/GLOBAL cases).
  • Forgetting the iterative nature of OR (always mention "review and feedback").

Final Visual Summary

mindmap
  root((Operational Research Summary))
    - Definition
      - Applied science for decision-making
    - Methodology
      - Problem → Model → Solve → Implement → Review
    - Tools
      - Linear Programming
      - Queuing Theory
      - Game Theory
    - Nepalese Examples
      - NTC Network Optimization
      - Daraz Inventory Management
    - Advantages
      - Cost-effective
      - Data-driven
    - Limitations
      - High computational cost
      - Model inaccuracies