Study guide
12.1 Demand Forecasting Methods and Forecast Error Tracking
FE Industrial and Systems Study Guide Exam
Track your progress as you read
Finished sections get a check mark.
1 Chapter 1: Exam Overview, CBT Strategy & Handbook Navigation
1.1 FE Industrial Exam Structure, CBT Logistics, and Pacing 1.2 Using the NCEES FE Reference Handbook and Approved Calculators
2 Chapter 2: Mathematics for Industrial Engineers
2.1 Analytic Geometry, Coordinate Systems, and Trigonometry 2.2 Calculus: Differential, Integral, and Series Approximations 2.3 Linear Algebra: Matrix Operations, Systems of Equations, and Vectors
3 Chapter 3: Engineering Sciences Core
3.1 Thermodynamics Fundamentals and Fluid Flow Principles 3.2 Statics, Dynamics, and Mechanics of Materials Core 3.3 Electrical Circuits, DC/AC Analysis, and Power in Industrial Systems
4 Chapter 4: Ethics and Professional Practice
4.1 NCEES Model Rules of Professional Conduct and Technical Society Ethics 4.2 Engineering Licensure, Public Welfare, and Professional Liability 4.3 Engineering Contracts, Intellectual Property, and Conflicts of Interest
5 Chapter 5: Engineering Economics
5.1 Time Value of Money, Discount Factors, and Equivalence 5.2 Evaluating Alternatives: Present Worth, Equivalent Annual Cost, and IRR 5.3 Incremental Rate of Return, Replacement Analysis, and Break-Even 5.4 Depreciation Methods, MACRS, Income Taxes, and Inflation
6 Chapter 6: Probability Foundations & Distributions
6.1 Basic Probability Concepts, Conditional Probability, and Bayes' Theorem 6.2 Discrete Probability Distributions 6.3 Continuous Probability Distributions
7 Chapter 7: Statistical Inference, Regression & Design of Experiments
7.1 Sampling Distributions, Central Limit Theorem, and Confidence Intervals 7.2 Hypothesis Testing: Means, Proportions, and Variances 7.3 Linear Regression, Correlation Analysis, and Design of Experiments (ANOVA)
8 Chapter 8: Data, Logic, Modeling & Linear Programming
8.1 Optimization Modeling: Decision Variables, Objectives, and Constraints 8.2 Linear Programming: Graphical Solutions and Simplex Fundamentals 8.3 Duality, Sensitivity Analysis, and Integer Linear Programming 8.4 Data, Logic Development, and Analytics
9 Chapter 9: Stochastic Models & Simulation
9.1 Queuing Theory: Little's Law, M/M/1, and Multi-Server Systems 9.2 Discrete-Time Markov Chains: Transition Matrices and Steady-State Probabilities 9.3 Discrete-Event Simulation Principles and Monte Carlo Modeling
10 Chapter 10: Engineering Management, Project Planning & Performance Measurement
10.1 Organizational Structure, Leadership, and Worker Motivation Theories 10.2 Project Management: WBS, Critical Path Method (CPM), PERT, and Crashing 10.3 Earned Value Management (EVM), Decision Trees, and Risk Preference 10.4 Performance Measurement: Productivity, KPIs, Wage Systems, and the Balanced Scorecard
11 Chapter 11: Manufacturing & Production Systems
11.1 Manufacturing Processes, Material Processing, and Tolerancing 11.2 Assembly Line Balancing, Takt Time, Cycle Time, and Throughput 11.3 Sustainable Manufacturing, Waste Minimization, and Life-Cycle Assessment
12 Chapter 12: Production Planning, Inventory Control & Sequencing
12.1 Demand Forecasting Methods and Forecast Error Tracking 12.2 Aggregate Planning, Master Production Scheduling (MPS), and MRP 12.3 Deterministic Inventory Models: EOQ, EPQ, and Safety Stock 12.4 Job Sequencing, Dispatching Rules, and Flow Shop Scheduling
13 Chapter 13: Lean Systems & Theory of Constraints
13.1 Lean Foundations: Value Stream Mapping, Seven Wastes, and 5S 13.2 Pull Systems, Kanban, Just-In-Time (JIT), and SMED Setup Reduction 13.3 Theory of Constraints: Bottleneck Identification and Drum-Buffer-Rope
14 Chapter 14: Facilities Planning & Material Handling
14.1 Facility Location Analysis: Qualitative Factor Rating and Centroid Method 14.2 Facility Layout Design and Flow Analysis 14.3 Material Handling Equipment, Warehousing, and AS/RS
15 Chapter 15: Supply Chain Design & Logistics
15.1 Supply Chain Network Architecture and Transportation Modes 15.2 Distribution Strategies, Cross-Docking, and Vehicle Routing 15.3 The Bullwhip Effect, Risk Pooling, and Supply Chain Coordination
16 Chapter 16: Human Factors & Cognitive Ergonomics
16.1 Cognitive Engineering, Human-Machine Interface, Displays, and Controls 16.2 Anthropometric Principles, Percentile Design, and Biomechanical Posture 16.3 Environmental Ergonomics: Industrial Noise, Lighting, and Heat Stress
17 Chapter 17: Physical Ergonomics & Workplace Safety
17.1 Physical Ergonomics: The NIOSH Manual Lifting Equation 17.2 Cumulative Trauma Disorders and Postural Assessment 17.3 Workplace Safety: OSHA Standards, Hazard Analysis, and Control Hierarchy
18 Chapter 18: Work Design & Measurement
18.1 Methods Engineering, Process Charting, and Principles of Motion Economy 18.2 Direct Stopwatch Time Study: Observed Time, Performance Rating, and Allowances 18.3 Predetermined Time Systems, Work Sampling, and Learning Curves
19 Chapter 19: Quality Management, Six Sigma & Analytical Tools
19.1 Quality Philosophies, Total Quality Management, and Cost of Quality 19.2 Six Sigma DMAIC Methodology and Quality Function Deployment 19.3 The Seven Basic Quality Tools, Root Cause Analysis, and Taguchi Loss Function
20 Chapter 20: Statistical Process Control
20.1 Statistical Process Control Principles, Variation, and Run Rules 20.2 Variable Control Charts: X-bar & R, X-bar & S, and Individual-Moving Range 20.3 Attribute Control Charts: Defectives (p, np) and Defects (c, u)
21 Chapter 21: Process Capability & Acceptance Sampling
21.1 Process Capability Indices and Defect Levels 21.2 Acceptance Sampling Concepts and Operating Characteristic Curves 21.3 Sampling Risks, AQL, LTPD, and Standard Sampling Plans
22 Chapter 22: Systems Engineering & Life-Cycle Management
22.1 The Systems Engineering V-Model, Stakeholder Requirements, and Functional Architecture 22.2 Configuration Management, Interface Control, and Verification & Validation 22.3 Life-Cycle Costing, Maintainability, and System Availability
23 Chapter 23: Reliability Engineering & Risk Analysis
23.1 Reliability Mathematics, Hazard Rates, and the Bathtub Curve 23.2 System Reliability Configurations: Series, Parallel, and k-out-of-n Systems 23.3 Failure Modes and Effects Analysis and Fault Tree Analysis
12.1 Demand Forecasting Methods and Forecast Error Tracking
Key Takeaways Explain This Topic AI
Quantitative forecasting relies on historical patterns (time series) or explanatory relationships (causal models), whereas qualitative forecasting (Delphi, executive opinion) is essential when historical data is absent or unrepresentative.
A time series decomposes into four structural components: horizontal baseline level, long-term trend, recurring seasonal cycles, and cyclical macroeconomic waves, surrounded by random irregular noise.
Simple exponential smoothing updates forecasts via F_{t+1} = F_t + α*(A_t - F_t); higher values of α increase responsiveness to recent demand shifts at the expense of noise dampening.
Forecast error metrics serve distinct diagnostic roles: MAD measures average error magnitude, MSE heavily penalizes large outliers, and MAPE provides scale-independent percentage evaluation.
The Tracking Signal (TS = CFE / MAD) identifies systematic forecasting bias; TS values exceeding ±4 to ±5 MADs signal that the model is out of statistical control, with σ ≈ 1.25 * MAD under normality.
Last updated: September 2026
Operations planning begins with an estimate of future customer demand. In industrial engineering, every downstream decision—from procurement quantities and production staffing to warehouse sizing and distribution logistics—depends fundamentally on the quality of the demand forecast. Because actual demand is inherently stochastic, industrial engineers must select appropriate mathematical forecasting techniques, track forecast errors over time, and establish statistical control limits to detect systemic bias.
1. Qualitative vs. Quantitative Forecasting Frameworks
Forecasting methodologies are broadly divided into two major classifications based on data availability and mathematical structure. Forecasting Methodologies
- ├── Qualitative (Subjective / Judgmental)
- │ ├── Delphi Method (Iterative, anonymous multi-round expert panel consensus)
- │ ├── Sales Force Composite (Aggregated field sales estimates)
- │ ├── Executive Opinion (High-level managerial consensus)
- │ └── Consumer Market Research (Surveys, focus groups, concept testing)
- └── Quantitative (Objective / Mathematical)
- ├── Time Series Models (Extrapolate historical chronological patterns)
- │ ├── Simple & Weighted Moving Averages
- │ ├── Simple Exponential Smoothing (SES)
- │ ├── Trend-Adjusted Exponential Smoothing (Holt's Method)
- │ └── Classical Seasonal Decomposition
- └── Causal / Associative Models (Explanatory mathematical relationships)
- ├── Simple Linear Regression (Y = β0 + β1*X + ε)
- └── Multiple Linear Regression & Econometric Models
- Qualitative Methods
Qualitative methods rely on human intuition, experience, and subjective judgment. They are critical when historical data is nonexistent (e.g., launching a revolutionary new product), when long-term strategic decisions (5–10 year horizon) render historical trends obsolete, or during severe market disruptions.
Delphi Method : An iterative group consensus technique where geographically separated experts answer questionnaires anonymously. A neutral facilitator summarizes responses and redistributes them through multiple rounds until consensus converges. Anonymity prevents dominant individuals from unduly influencing the group.
Sales Force Composite : Regional sales representatives estimate demand in their territories. While close to customers, estimates often suffer from optimism bias or deliberate underestimation if sales quotas are tied to forecasts.
Executive Opinion : High-level cross-functional managers (finance, operations, marketing) develop joint estimates. Fast and strategic, but vulnerable to corporate hierarchy bias.
Quantitative Methods
Quantitative methods require historical numerical data and assume that past mathematical patterns will persist into the future. They are objective, reproducible, and easily automated within Enterprise Resource Planning (ERP) databases.
Attribute Qualitative Methods Time Series Models Causal / Associative Models
Data Requirement Little to no historical numerical data Historical observations of variable Y t Y_t Y t Historical data of Y t Y_t Y t and predictors X i X_i X i
Planning Horizon Long-range strategic (1–10 years) Short-to-medium range (days to months) Medium-to-long range (months to years)
Mathematical Basis Consensus, intuition, surveys Statistical smoothing, autocorrelation Ordinary Least Squares regression, correlation
Primary Strength Captures paradigm shifts and new products Low computation, reacts to patterns Explains why demand changes via drivers
2. Components of a Time Series
A time series is a sequence of observations recorded at regular, uniform time intervals (hourly, daily, weekly, monthly, quarterly). Classical time series analysis assumes the observed historical demand Y t Y_t Y t is composed of four distinct underlying components:
Horizontal Level ( L L L ) : The baseline average value around which demand fluctuates when no trends or seasonal cycles exist.
Trend ( T T T ) : The long-term persistent upward or downward movement in the series over an extended timeframe, driven by population growth, market adoption, or technological obsolescence.
Seasonality ( S S S ) : A regular, repeating pattern of peaks and valleys that recurs over a known, fixed duration (e.g., weekly retail spikes on Saturdays, quarterly agricultural cycles, hourly power consumption).
Cyclical ( C C C ) : Multi-year wave-like oscillations tied to macroeconomic expansions and recessions. Unlike seasonality, cyclical movements vary in length (typically 2–10 years) and amplitude, making them difficult to model with simple time series.
Random Variation / Irregular Noise ( ϵ \epsilon ϵ ) : Uncorrelated, unpredictable residual noise caused by weather anomalies, equipment breakdowns, or chance occurrences ( E [ ϵ ] = 0 E[\epsilon] = 0 E [ ϵ ] = 0 ).
Mathematical Formulation: Additive vs. Multiplicative
Additive Model : Assumes components are independent and expressed in physical demand units: Y t = L + T t + S t + C t + ϵ t Y_t = L + T_t + S_t + C_t + \epsilon_t Y t = L + T t + S t + C t + ϵ t When to use : Seasonal oscillations remain constant in magnitude regardless of overall demand level.
Multiplicative Model : Assumes seasonal variations scale proportionally with the baseline level and trend: Y t = ( L + T t ) × S t × C t × ϵ t Y_t = (L + T_t) \times S_t \times C_t \times \epsilon_t Y t = ( L + T t ) × S t × C t × ϵ t When to use : Seasonal swings amplify as baseline volume expands (standard industrial behavior).
3. Moving Average Forecasting Methods
Moving average models smooth out short-term random noise by calculating the arithmetic mean of the most recent historical observations.
Simple Moving Average (SMA)
For a moving average of order k k k (spanning k k k historical periods), the forecast for period t + 1 t+1 t + 1 is: F t + 1 = 1 k ∑ i = 0 k − 1 A t − i = A t + A t − 1 + A t − 2 + ⋯ + A t − k + 1 k F_{t+1} = \frac{1}{k} \sum_{i=0}^{k-1} A_{t-i} = \frac{A_t + A_{t-1} + A_{t-2} + \dots + A_{t-k+1}}{k} F t + 1 = k 1 i = 0 ∑ k − 1 A t − i = k A t + A t − 1 + A t − 2 + ⋯ + A t − k + 1
where A t A_t A t represents the actual observed demand in period t t t .
Parameter Selection Trade-off :
Large k k k (e.g., k = 12 k = 12 k = 12 ) : Maximum smoothing stability. Highly effective at filtering out severe white noise, but sluggish and slow to respond to genuine structural shifts or trends.
Small k k k (e.g., k = 3 k = 3 k = 3 ) : High responsiveness. Rapidly tracks recent demand shifts, but susceptible to over-reacting to transient random fluctuations.
Fundamental Limitation : All moving averages lag behind an upward or downward trend. In an upward trend, SMA will consistently underestimate actual demand ( A t > F t A_t > F_t A t > F t );
in a downward trend, SMA will consistently overestimate demand ( A t < F t A_t < F_t A t < F t ).
Weighted Moving Average (WMA)
To improve responsiveness without discarding older data entirely, unequal weights w i w_i w i are assigned to past periods, typically giving the highest weight to the most recent observation: F t + 1 = ∑ i = 1 k w i A t − i + 1 = w 1 A t + w 2 A t − 1 + ⋯ + w k A t − k + 1 F_{t+1} = \sum_{i=1}^k w_i A_{t-i+1} = w_1 A_t + w_2 A_{t-1} + \dots + w_k A_{t-k+1} F t + 1 = i = 1 ∑ k w i A t − i + 1 = w 1 A t + w 2 A t − 1 + ⋯ + w k A t − k + 1 Subject to: ∑ i = 1 k w i = 1.0 , w i ≥ 0 \text{Subject to: } \sum_{i=1}^k w_i = 1.0, \quad w_i \ge 0 Subject to: i = 1 ∑ k w i = 1.0 , w i ≥ 0
4. Exponential Smoothing Techniques
Exponential smoothing is the most widely implemented quantitative forecasting method in industry because it requires storing only two numbers in memory: the most recent actual demand ( A t A_t A t ) and the most recent forecast ( F t F_t F t ). Older observations are weighted with geometrically decaying importance.
Simple Exponential Smoothing (SES)
The forecast for period t + 1 t+1 t + 1 is formulated as a convex combination of actual demand in period t t t and the forecast made for period t t t : F t + 1 = α A t + ( 1 − α ) F t F_{t+1} = \alpha A_t + (1 - \alpha) F_t F t + 1 = α A t + ( 1 − α ) F t
Rearranging into the intuitive error-correction form : F t + 1 = F t + α ( A t − F t ) = F t + α e t F_{t+1} = F_t + \alpha (A_t - F_t) = F_t + \alpha e_t F t + 1 = F t + α ( A t − F t ) = F t + α e t
where α \alpha α is the smoothing constant ( 0 ≤ α ≤ 1 0 \le \alpha \le 1 0 ≤ α ≤ 1 ) and e t = A t − F t e_t = A_t - F_t e t = A t − F t is the forecast error in period t t t . Exponential Decay of Weights
F_{t+1} = α*A_t + α*(1-α)*A_{t-1} + α*(1-α)^2*A_{t-2} + α*(1-α)^3*A_{t-3} + ...
- Weights: [α] [α(1-α)] [α(1-α)^2] [α(1-α)^3]
- Sum of infinite weights = α / [1 - (1 - α)] = 1.0
- The Role of the Smoothing Constant α \alpha α
High α \alpha α (e.g., α = 0.7 − 0.9 \alpha = 0.7 - 0.9 α = 0.7 − 0.9 ) : Puts heavy weight on the latest observation. Forecast reacts rapidly to recent changes. Equivalent to a small-period moving average.
Low α \alpha α (e.g., α = 0.1 − 0.2 \alpha = 0.1 - 0.2 α = 0.1 − 0.2 ) : Puts heavy weight on past forecast history. Filters out noise, providing smooth, stable forecasts.
Equivalent Moving Average Span ( k k k ) : The approximate correspondence between smoothing constant α \alpha α and moving average span k k k is given by: α ≈ 2 k + 1 ⟺ k ≈ 2 α − 1 \alpha \approx \frac{2}{k + 1} \quad \Longleftrightarrow \quad k \approx \frac{2}{\alpha} - 1 α ≈ k + 1 2 ⟺ k ≈ α 2 − 1 For example, α = 0.2 \alpha = 0.2 α = 0.2 behaves similarly to a k = 2 0.2 − 1 = 9 k = \frac{2}{0.2} - 1 = 9 k = 0.2 2 − 1 = 9 -period moving average.
Trend-Adjusted Exponential Smoothing (Holt's Method)
Just like moving averages, simple exponential smoothing exhibits systematic lag when demand contains a persistent trend. Holt's two-parameter exponential smoothing corrects for trend by tracking both a smoothed level ( L t L_t L t ) and a smoothed trend rate ( T t T_t T t ):
Update Base Level ( L t L_t L t ) : L t = α A t + ( 1 − α ) ( L t − 1 + T t − 1 ) L_t = \alpha A_t + (1 - \alpha)(L_{t-1} + T_{t-1}) L t = α A t + ( 1 − α ) ( L t − 1 + T t − 1 )
Update Trend Estimate ( T t T_t T t ) : T t = β ( L t − L t − 1 ) + ( 1 − β ) T t − 1 T_t = \beta (L_t - L_{t-1}) + (1 - \beta) T_{t-1} T t = β ( L t − L t − 1 ) + ( 1 − β ) T t − 1
Forecast p p p Periods Ahead ( F t + p F_{t+p} F t + p ) : F t + p = L t + p ⋅ T t F_{t+p} = L_t + p \cdot T_t F t + p = L t + p ⋅ T t
where α \alpha α is the level smoothing constant and β \beta β is the trend smoothing constant ( 0 ≤ α , β ≤ 1 0 \le \alpha, \beta \le 1 0 ≤ α , β ≤ 1 ).
5. Seasonal Indices and Classical Decomposition
When demand exhibits cyclical peaks and troughs at regular intervals within a year (or week), industrial engineers employ seasonal decomposition.
Calculating Seasonal Indices (Multiplicative Approach)
A Seasonal Index ( S I i SI_i S I i ) represents the ratio of average demand during season i i i to the average demand across all seasons: S I i = D ˉ i D ˉ overall SI_i = \frac{\bar{D}_i}{\bar{D}_{\text{overall}}} S I i = D ˉ overall D ˉ i ∑ i = 1 m S I i = m \sum_{i=1}^m SI_i = m i = 1 ∑ m S I i = m
where m m m is the number of distinct seasons in a full cycle (e.g., m = 4 m = 4 m = 4 for quarters, m = 12 m = 12 m = 12 for months, m = 7 m = 7 m = 7 for days of the week).
S I i > 1.0 SI_i > 1.0 S I i > 1.0 : Demand in period i i i is above the annual average (e.g., S I = 1.25 SI = 1.25 S I = 1.25 indicates demand is 25% above average).
S I i < 1.0 SI_i < 1.0 S I i < 1.0 : Demand in period i i i is below average (e.g., S I = 0.80 SI = 0.80 S I = 0.80 indicates demand is 20% below average).
The Forecasting Procedure with Seasonality
Deseasonalize Historical Data : Divide actual demand by the corresponding seasonal index to uncover the pure underlying trend and level: Deseasonalized Demand t = A t S I t \text{Deseasonalized Demand}_t = \frac{A_t}{SI_t} Deseasonalized Demand t = S I t A t
Estimate Trend / Baseline : Fit a linear regression or moving average to the deseasonalized data: Y ^ t = a + b t \hat{Y}_t = a + b t Y ^ t = a + b t .
Project Future Unadjusted Forecast : Calculate future baseline values Y ^ t + p \hat{Y}_{t+p} Y ^ t + p .
Reseasonalize : Multiply the baseline projection by the target period's seasonal index: F t + p = Y ^ t + p × S I t + p F_{t+p} = \hat{Y}_{t+p} \times SI_{t+p} F t + p = Y ^ t + p × S I t + p
6. Measuring Forecast Error
Forecast error in period t t t is universally defined under NCEES standards as actual observed demand minus forecasted demand: ≠ t = A t − F t \ne_t = A_t - F_t = t = A t − F t
Exam Convention Alert : Always subtract Forecast from Actual ( A t − F t A_t - F_t A t − F t ). A positive error ( e t > 0 e_t > 0 e t > 0 ) means actual demand was higher than forecasted (underforecasting → \to → stockout risk). A negative error ( e t < 0 e_t < 0 e t < 0 ) means actual demand was lower than forecasted (overforecasting → \to → excess inventory risk).
Core Error Metrics
Metric Mathematical Formulation Characteristic & Operational Purpose
Mean Forecast Error (MFE / Bias) MFE = 1 n ∑ t = 1 n e t \text{MFE} = \frac{1}{n} \sum_{t=1}^n e_t MFE = n 1 ∑ t = 1 n e t Measures directional bias. Unbiased models have MFE ≈ 0 \text{MFE} \approx 0 MFE ≈ 0 . Positive ⟹ \implies ⟹ underforecasting; negative ⟹ \implies ⟹ overforecasting. Does not measure error magnitude because positive and negative errors cancel out.
Mean Absolute Deviation (MAD) MAD = 1 n ∑ t = 1 n ∣ e t ∣ \text{MAD} = \frac{1}{n} \sum_{t=1}^n \vert e_t\vert MAD = n 1 ∑ t = 1 n ∣ e t ∣ Measures average absolute error magnitude in native demand units. Treats positive and negative errors identically without canceling. Used directly to size safety stock.
Mean Squared Error (MSE) MSE = 1 n ∑ t = 1 n e t 2 \text{MSE} = \frac{1}{n} \sum_{t=1}^n e_t^2 MSE = n 1 ∑ t = 1 n e t 2 Penalizes large errors disproportionately due to squaring. Mathematically analogous to error variance (sample variance when bias is zero).
Root Mean Squared Error (RMSE) RMSE = MSE \text{RMSE} = \sqrt{\text{MSE}} RMSE = MSE Converts squared error penalty back into native units of demand. Expresses standard error of the estimate.
Mean Absolute Percentage Error (MAPE) MAPE = 100 % n ∑ t = 1 n ∣ e t A t ∣ \text{MAPE} = \frac{100\%}{n} \sum_{t=1}^n \left\vert\frac{e_t}{A_t} \right\vert MAPE = n 100% ∑ t = 1 n A t e t Measures error relative to actual demand magnitude. Dimensionless percentage; allows direct comparisons across products with vastly different sales volumes. Undefined if A t = 0 A_t = 0 A t = 0 .
7. Tracking Signals and Forecast Control Limits
A forecasting model may have a low MAD, yet still be systematically biased if it continually under- or over-predicts demand. Industrial engineers monitor model validity using the Tracking Signal (TS) .
Cumulative Forecast Error (CFE)
The algebraic sum of all historical errors, also called the Running Sum of Forecast Errors (RSFE): CFE n = ∑ t = 1 n e t = ∑ t = 1 n ( A t − F t ) \text{CFE}_n = \sum_{t=1}^n e_t = \sum_{t=1}^n (A_t - F_t) CFE n = t = 1 ∑ n e t = t = 1 ∑ n ( A t − F t )
Tracking Signal Formulation
The Tracking Signal divides the cumulative algebraic error by the current Mean Absolute Deviation: TS t = CFE t MAD t = ∑ i = 1 t e i 1 t ∑ i = 1 t ∣ e i ∣ \text{TS}_t = \frac{\text{CFE}_t}{\text{MAD}_t} = \frac{\sum_{i=1}^t e_i}{\frac{1}{t}\sum_{i=1}^t |e_i|} TS t = MAD t CFE t = t 1 ∑ i = 1 t ∣ e i ∣ ∑ i = 1 t e i Tracking Signal Control Limits
- +5 MAD ─────────────────────────────── Upper Control Limit (Severe Underforecasting)
- +4 MAD ------------------------------- Warning Threshold
- 0 MAD ═══════════════════════════════ Ideal Unbiased Performance (CFE = 0)
- 4 MAD ------------------------------- Warning Threshold
- 5 MAD ─────────────────────────────── Lower Control Limit (Severe Overforecasting)
- Statistical Control Limits on Tracking Signal
In-Control Range : A tracking signal between − 4 -4 − 4 and + 4 +4 + 4 (or up to ± 5 \pm 5 ± 5 in some industries) indicates the forecasting model is in statistical control with random, zero-mean error fluctuations.
Out-of-Control Condition : If ∣ TS ∣ > 4 |\text{TS}| > 4 ∣ TS ∣ > 4 (or 5 5 5 ), the model is declared biased and out of control:
TS > + 4 \text{TS} > +4 TS > + 4 : Consistent underforecasting (actuals repeatedly exceed forecasts; stockouts imminent). Model needs an upward level/trend adjustment.
TS < − 4 \text{TS} < -4 TS < − 4 : Consistent overforecasting (forecasts repeatedly exceed actuals; inventory pile-up imminent). Model needs a downward level/trend adjustment.
Relationship Between MAD and Standard Deviation ( σ \sigma σ )
If forecast errors are approximately normally distributed with zero mean, the standard deviation of forecast errors ( σ \sigma σ ) is mathematically related to the Mean Absolute Deviation by: σ = π 2 ⋅ MAD ≈ 1.2533 ⋅ MAD ≈ 1.25 ⋅ MAD \sigma = \sqrt{\frac{\pi}{2}} \cdot \text{MAD} \approx 1.2533 \cdot \text{MAD} \approx 1.25 \cdot \text{MAD} σ = 2 π ⋅ MAD ≈ 1.2533 ⋅ MAD ≈ 1.25 ⋅ MAD MAD = 2 π ⋅ σ ≈ 0.7979 ⋅ σ ≈ 0.80 ⋅ σ \text{MAD} = \sqrt{\frac{2}{\pi}} \cdot \sigma \approx 0.7979 \cdot \sigma \approx 0.80 \cdot \sigma MAD = π 2 ⋅ σ ≈ 0.7979 ⋅ σ ≈ 0.80 ⋅ σ
This fundamental identity allows engineers to translate a calculated MAD directly into standard deviations for setting safety stock and reorder points without calculating individual squared residuals.
- 8. Step-by-Step Worked Engineering Calculations
- Worked Example 12.1.1: Exponential Smoothing, Error Metrics, and Tracking Signal
Problem : A manufacturing plant monitors monthly demand for an industrial electric motor. Demand over the past 4 months is recorded below. At the beginning of Month 1, the forecast was F 1 = 110 F_1 = 110 F 1 = 110 units. The company uses simple exponential smoothing with a smoothing constant α = 0.30 \alpha = 0.30 α = 0.30 .
- Month ( t t t ) Actual Demand ( A t A_t A t )
- 1 120
- 2 130
- 3 110
- 4 125
Calculate the forecasts F 2 , F 3 , F 4 F_2, F_3, F_4 F 2 , F 3 , F 4 , and the projected forecast for Month 5 ( F 5 F_5 F 5 ).
Compute the forecast errors e t e_t e t , absolute errors ∣ e t ∣ |e_t| ∣ e t ∣ , squared errors e t 2 e_t^2 e t 2 , and absolute percentage errors ∣ e t / A t ∣ |e_t / A_t| ∣ e t / A t ∣ for months 1 through 4.
Determine the MAD, MSE, MAPE, and Cumulative Forecast Error (CFE) across the 4 months.
Calculate the Tracking Signal at the end of Month 4 and evaluate whether the model is in statistical control.
Estimate the standard deviation of forecast errors ( σ \sigma σ ).
Solution :
Step 1: Compute Forecasts Using F t + 1 = F t + α ( A t − F t ) F_{t+1} = F_t + \alpha (A_t - F_t) F t + 1 = F t + α ( A t − F t )
Month 1: Given F 1 = 110.00 F_1 = 110.00 F 1 = 110.00 . Actual A 1 = 120 A_1 = 120 A 1 = 120 . e 1 = A 1 − F 1 = 120 − 110 = + 10.00 e_1 = A_1 - F_1 = 120 - 110 = +10.00 e 1 = A 1 − F 1 = 120 − 110 = + 10.00
Month 2: F 2 = 110.00 + 0.30 ( 120 − 110 ) = 110.00 + 3.00 = 113.00 F_2 = 110.00 + 0.30(120 - 110) = 110.00 + 3.00 = 113.00 F 2 = 110.00 + 0.30 ( 120 − 110 ) = 110.00 + 3.00 = 113.00 e 2 = A 2 − F 2 = 130 − 113.00 = + 17.00 e_2 = A_2 - F_2 = 130 - 113.00 = +17.00 e 2 = A 2 − F 2 = 130 − 113.00 = + 17.00
Month 3: F 3 = 113.00 + 0.30 ( 130 − 113.00 ) = 113.00 + 5.10 = 118.10 F_3 = 113.00 + 0.30(130 - 113.00) = 113.00 + 5.10 = 118.10 F 3 = 113.00 + 0.30 ( 130 − 113.00 ) = 113.00 + 5.10 = 118.10 e 3 = A 3 − F 3 = 110 − 118.10 = − 8.10 e_3 = A_3 - F_3 = 110 - 118.10 = -8.10 e 3 = A 3 − F 3 = 110 − 118.10 = − 8.10
Month 4: F 4 = 118.10 + 0.30 ( 110 − 118.10 ) = 118.10 − 2.43 = 115.67 F_4 = 118.10 + 0.30(110 - 118.10) = 118.10 - 2.43 = 115.67 F 4 = 118.10 + 0.30 ( 110 − 118.10 ) = 118.10 − 2.43 = 115.67 e 4 = A 4 − F 4 = 125 − 115.67 = + 9.33 e_4 = A_4 - F_4 = 125 - 115.67 = +9.33 e 4 = A 4 − F 4 = 125 − 115.67 = + 9.33
Projected Forecast for Month 5 ( F 5 F_5 F 5 ): F 5 = 115.67 + 0.30 ( 125 − 115.67 ) = 115.67 + 0.30 ( 9.33 ) = 115.67 + 2.80 = 118.47 units F_5 = 115.67 + 0.30(125 - 115.67) = 115.67 + 0.30(9.33) = 115.67 + 2.80 = 118.47\text{ units} F 5 = 115.67 + 0.30 ( 125 − 115.67 ) = 115.67 + 0.30 ( 9.33 ) = 115.67 + 2.80 = 118.47 units
Step 2: Construct the Comprehensive Error Tracking Table
t t t A t A_t A t F t F_t F t e t = A t − F t e_t = A_t - F_t e t = A t − F t ∣ e t ∣ \vert e_t\vert ∣ e t ∣ e t 2 e_t^2 e t 2 ∣ e t / A t ∣ × 100 % \vert e_t / A_t\vert\times 100\% ∣ e t / A t ∣ × 100% Cumulative e t e_t e t (CFE)
- 1 120 110.00 + 10.00 +10.00 + 10.00 10.00 100.00 8.33 % 8.33\% 8.33% + 10.00 +10.00 + 10.00
- 2 130 113.00 + 17.00 +17.00 + 17.00 17.00 289.00 13.08 % 13.08\% 13.08% + 27.00 +27.00 + 27.00
- 3 110 118.10 − 8.10 -8.10 − 8.10 8.10 65.61 7.36 % 7.36\% 7.36% + 18.90 +18.90 + 18.90
- 4 125 115.67 + 9.33 +9.33 + 9.33 9.33 87.05 7.46 % 7.46\% 7.46% + 28.23 +28.23 + 28.23
- Sum — — + 28.23 +28.23 + 28.23 44.43 541.66 36.23 % 36.23\% 36.23% —
- Step 3: Evaluate Error Metrics
MAD : MAD = ∑ ∣ e t ∣ n = 44.43 4 = 11.11 units \text{MAD} = \frac{\sum |e_t|}{n} = \frac{44.43}{4} = 11.11\text{ units} MAD = n ∑ ∣ e t ∣ = 4 44.43 = 11.11 units
MSE : MSE = ∑ e t 2 n = 541.66 4 = 135.42 units 2 \text{MSE} = \frac{\sum e_t^2}{n} = \frac{541.66}{4} = 135.42\text{ units}^2 MSE = n ∑ e t 2 = 4 541.66 = 135.42 units 2
RMSE : RMSE = 135.42 = 11.64 units \text{RMSE} = \sqrt{135.42} = 11.64\text{ units} RMSE = 135.42 = 11.64 units
MAPE : MAPE = 36.23 % 4 = 9.06 % \text{MAPE} = \frac{36.23\%}{4} = 9.06\% MAPE = 4 36.23% = 9.06%
CFE (Running Sum of Forecast Errors at Month 4): CFE 4 = + 10.00 + 17.00 − 8.10 + 9.33 = + 28.23 units \text{CFE}_4 = +10.00 + 17.00 - 8.10 + 9.33 = +28.23\text{ units} CFE 4 = + 10.00 + 17.00 − 8.10 + 9.33 = + 28.23 units
Step 4: Tracking Signal Evaluation TS 4 = CFE 4 MAD 4 = + 28.23 11.11 = + 2.54 \text{TS}_4 = \frac{\text{CFE}_4}{\text{MAD}_4} = \frac{+28.23}{11.11} = +2.54 TS 4 = MAD 4 CFE 4 = 11.11 + 28.23 = + 2.54
Evaluation : Because ∣ TS 4 ∣ = 2.54 ≤ 4.0 |\text{TS}_4| = 2.54 \le 4.0 ∣ TS 4 ∣ = 2.54 ≤ 4.0 , the forecasting model is in statistical control . However, the positive value ( + 2.54 +2.54 + 2.54 ) reveals an underlying tendency toward underforecasting, warning the engineer that demand may be trending upward faster than α = 0.30 \alpha = 0.30 α = 0.30 can accommodate.
Step 5: Estimate Standard Deviation of Forecast Error σ ≈ 1.25 × MAD = 1.25 × 11.11 = 13.89 units \sigma \approx 1.25 \times \text{MAD} = 1.25 \times 11.11 = 13.89\text{ units} σ ≈ 1.25 × MAD = 1.25 × 11.11 = 13.89 units
Worked Example 12.1.2: Seasonal Index Decomposition and Forecasting
Problem : A manufacturer of commercial HVAC units experiences strong seasonal demand. Historical quarterly sales over the past two years are presented below. Corporate sales forecasts project total annual demand for Year 3 at 1,440 units.
- Quarter Year 1 Sales Year 2 Sales
- Q1 220 260
- Q2 380 420
- Q3 460 500
- Q4 140 180
Calculate the seasonal index for each of the four quarters using the multiplicative decomposition method.
Determine the quarterly demand forecasts for Year 3.
Solution :
Step 1: Determine Average Demand by Quarter and Overall Average
Average Q1 Demand: D ˉ 1 = 220 + 260 2 = 240 units \bar{D}_1 = \frac{220 + 260}{2} = 240\text{ units} D ˉ 1 = 2 220 + 260 = 240 units
Average Q2 Demand: D ˉ 2 = 380 + 420 2 = 400 units \bar{D}_2 = \frac{380 + 420}{2} = 400\text{ units} D ˉ 2 = 2 380 + 420 = 400 units
Average Q3 Demand: D ˉ 3 = 460 + 500 2 = 480 units \bar{D}_3 = \frac{460 + 500}{2} = 480\text{ units} D ˉ 3 = 2 460 + 500 = 480 units
Average Q4 Demand: D ˉ 4 = 140 + 180 2 = 160 units \bar{D}_4 = \frac{140 + 180}{2} = 160\text{ units} D ˉ 4 = 2 140 + 180 = 160 units
Total Average Annual Cycle Demand: 240 + 400 + 480 + 160 = 1 , 280 units 240 + 400 + 480 + 160 = 1,280\text{ units} 240 + 400 + 480 + 160 = 1 , 280 units
Overall Average Quarterly Baseline ( D ˉ overall \bar{D}_{\text{overall}} D ˉ overall ): D ˉ overall = 1 , 280 4 = 320 units/quarter \bar{D}_{\text{overall}} = \frac{1,280}{4} = 320\text{ units/quarter} D ˉ overall = 4 1 , 280 = 320 units/quarter
Step 2: Calculate Seasonal Indices ( S I i = D ˉ i / D ˉ overall SI_i = \bar{D}_i / \bar{D}_{\text{overall}} S I i = D ˉ i / D ˉ overall )
- S I 1 = 240 320 = 0.75 SI_1 = \frac{240}{320} = 0.75 S I 1 = 320 240 = 0.75
- S I 2 = 400 320 = 1.25 SI_2 = \frac{400}{320} = 1.25 S I 2 = 320 400 = 1.25
- S I 3 = 480 320 = 1.50 SI_3 = \frac{480}{320} = 1.50 S I 3 = 320 480 = 1.50
- S I 4 = 160 320 = 0.50 SI_4 = \frac{160}{320} = 0.50 S I 4 = 320 160 = 0.50
Verification : ∑ S I i = 0.75 + 1.25 + 1.50 + 0.50 = 4.00 = m \sum SI_i = 0.75 + 1.25 + 1.50 + 0.50 = 4.00 = m ∑ S I i = 0.75 + 1.25 + 1.50 + 0.50 = 4.00 = m . (Matches cycle length).
Step 3: Forecast Year 3 Quarterly Demand
Year 3 expected baseline quarterly demand without seasonality: Baseline Quarter = 1 , 440 4 = 360 units/quarter \text{Baseline Quarter} = \frac{1,440}{4} = 360\text{ units/quarter} Baseline Quarter = 4 1 , 440 = 360 units/quarter
Multiply baseline by respective seasonal indices ( F i = 360 × S I i F_i = 360 \times SI_i F i = 360 × S I i ):
Q1 Forecast : F 1 = 360 × 0.75 = 270 units F_1 = 360 \times 0.75 = 270\text{ units} F 1 = 360 × 0.75 = 270 units
Q2 Forecast : F 2 = 360 × 1.25 = 450 units F_2 = 360 \times 1.25 = 450\text{ units} F 2 = 360 × 1.25 = 450 units
Q3 Forecast : F 3 = 360 × 1.50 = 540 units F_3 = 360 \times 1.50 = 540\text{ units} F 3 = 360 × 1.50 = 540 units
Q4 Forecast : F 4 = 360 × 0.50 = 180 units F_4 = 360 \times 0.50 = 180\text{ units} F 4 = 360 × 0.50 = 180 units
Check : 270 + 450 + 540 + 180 = 1 , 440 units 270 + 450 + 540 + 180 = 1,440\text{ units} 270 + 450 + 540 + 180 = 1 , 440 units .
9. NCEES Reference Handbook Tips & Realistic Exam Traps
Error Sign Convention Trap : NCEES defines error as e t = A t − F t e_t = A_t - F_t e t = A t − F t (Actual minus Forecast). If a problem gives you Actual = 100 and Forecast = 120, e t = − 20 e_t = -20 e t = − 20 . A negative error indicates overforecasting. Do not write F t − A t F_t - A_t F t − A t .
Smoothing Constant Responsiveness : Remember that higher α \alpha α means more responsive to recent demand changes (giving weight α \alpha α to A t A_t A t ), while lower α \alpha α provides smoother, more stable forecasts. On the exam, questions asking "which parameter value should be chosen if the demand process experiences rapid, permanent structural shifts?" require selecting the highest α \alpha α (e.g., 0.7 to 0.9).
The MAD to σ \sigma σ Conversion : Memorize σ ≈ 1.25 × MAD \sigma \approx 1.25 \times \text{MAD} σ ≈ 1.25 × MAD . A common exam trap is to divide MAD by 1.25 (or multiply by 0.80) to find σ \sigma σ . The standard deviation of a normal distribution is always larger than its Mean Absolute Deviation ( σ = 1.2533 × MAD \sigma = 1.2533 \times \text{MAD} σ = 1.2533 × MAD , while MAD = 0.7979 × σ \text{MAD} = 0.7979 \times \sigma MAD = 0.7979 × σ ).
Tracking Signal Denominator : The denominator of Tracking Signal is MAD , NOT MSE, RMSE, or σ \sigma σ . Tracking Signal has no units because CFE and MAD share identical units (both are in units of physical product).
Deseasonalizing vs. Reseasonalizing : When eliminating seasonality from actual data, you divide by the seasonal index ( A t / S I A_t / SI A t / S I ). When creating the final operational forecast from a baseline trend, you multiply by the seasonal index ( F t = Baseline × S I F_t = \text{Baseline} \times SI F t = Baseline × S I ).