Portfolio Optimization and Resource Allocation

Introduction

Product and portfolio management is fundamentally a problem of choice under constraint. Most organizations have more opportunities than they can adequately fund, staff, launch, or govern. The portfolio may contain established products, incremental improvements, platform investments, experiments, partnerships, and high-uncertainty innovations. Each competes for scarce capital, engineering capacity, data, managerial attention, channel access, and customer tolerance for change.

Portfolio optimization is therefore not equivalent to selecting the products with the highest individual forecast return. It is the disciplined design of a set of initiatives that collectively advances strategy while respecting financial, operational, technological, regulatory, and organizational constraints. A well-optimized portfolio balances short-term cash generation with long-term growth, exploits existing capabilities while exploring new ones, and makes uncertainty visible rather than disguising it in precise-looking forecasts.

This chapter develops an integrated approach to portfolio prioritization. It combines strategic scoring, financial evaluation, scenario analysis, real-options reasoning, risk assessment, dependency management, and constrained optimization. The emphasis is managerial: quantitative models support judgment, but they do not replace strategic interpretation, governance, or ethical accountability.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish product-level attractiveness from portfolio-level value.
2. Construct and critique weighted scoring models.
3. Apply NPV, profitability index, payback, and expected-value logic appropriately.
4. Incorporate uncertainty through scenarios, sensitivity analysis, and risk-adjusted evaluation.
5. Explain how real options affect innovation investment decisions.
6. Formulate resource-allocation problems with budget, capacity, and dependency constraints.
7. Interpret integer and linear optimization outputs without treating them as automatic decisions.
8. Design governance processes that manage trade-offs, bias, and changing information.

1. The Portfolio as a Strategic System

A product portfolio is a set of offerings and initiatives managed as an interdependent system. It can include:

• mature products that generate cash;
• growth products requiring expansion investment;
• declining products requiring harvesting or exit decisions;
• innovation concepts and minimum viable products;
• shared platforms, data assets, and enabling technologies;
• regulatory, quality, and resilience investments.

The value of an initiative depends partly on its relationship with other initiatives. A product may share a platform with another product, cannibalize an existing offering, unlock a new channel, or consume a scarce specialist team. Consequently, the relevant question is not simply “Which project has the highest return?” but “Which combination creates the greatest strategically acceptable value?”

Portfolio objectives

Portfolio objectives should be stated before scoring initiatives. Typical objectives include:

• revenue or contribution growth;
• economic profit and cash generation;
• customer retention and lifetime value;
• strategic entry into a market or capability;
• brand coherence and differentiation;
• resilience, compliance, and risk reduction;
• learning and option creation.

These objectives may conflict. A compliance initiative may have little direct revenue but prevent severe downside. A platform investment may have negative standalone NPV but enable several profitable products. A premium product may increase brand equity while reducing short-term volume. Portfolio management makes such trade-offs explicit.

Balance dimensions

Four balancing dimensions are especially useful:

Balance does not mean equal allocation. It means that the allocation is intentional and consistent with risk-bearing capacity and strategy.

2. From Strategic Criteria to Comparable Priorities

Weighted scoring models

A weighted scoring model converts multiple criteria into a transparent priority index. For initiative i, the basic additive model is:

where Sᵢ is the score of initiative i; wⱼ is the weight of criterion j; rᵢⱼ is the normalized rating of initiative i on criterion j; and m is the number of criteria. Usually, weights satisfy:

The ratings must use a consistent scale, such as 0–10, and higher values must always represent greater desirability. If a criterion is a cost or risk, its scale must be reversed or transformed before aggregation.

Assumptions and limitations. The additive model assumes that criteria can be meaningfully traded off and that the weights represent decision-maker preferences. It also assumes that criteria are sufficiently distinct. Double-counting “market growth,” “revenue potential,” and “market attractiveness” can create false confidence. The model is best used for structured comparison, not as a claim that strategic value is precisely measurable.

Example of a scoring model

Suppose a firm uses four criteria: strategic fit (weight 0.30), financial potential (0.30), customer value (0.20), and feasibility (0.20). An initiative receives ratings of 9, 7, 8, and 5 respectively on a 0–10 scale. Its score is:

The score is 7.4 out of 10. It does not mean that the initiative will generate 7.4 units of profit. It means that, under the stated criteria and weights, the initiative has a relatively strong qualitative priority.

Improving scoring-model reliability

Professional practice should include:

1. Explicit criterion definitions. “Strategic fit” should identify the strategy elements being assessed.
2. Evidence-based ratings. Ratings should refer to customer research, financial estimates, experiments, or expert evidence.
3. Independent scoring. Independent assessments reduce conformity and political influence.
4. Calibration sessions. Reviewers should discuss why a rating of 8 differs from 6.
5. Sensitivity testing. Recalculate rankings under plausible changes in weights and ratings.
6. Separate gates from preferences. Legal compliance and minimum technical feasibility may be non-negotiable gates rather than weighted criteria.

A useful extension is a two-stage process: first eliminate initiatives that fail mandatory requirements; then rank the remaining initiatives using a weighted model.

3. Financial Evaluation of Products and Initiatives

Strategic scoring should be complemented by financial analysis. Different metrics answer different questions and should not be treated as interchangeable.

Net present value

Net present value measures the present value of expected future cash flows less the initial investment:

Here, I₀ is the initial cash outflow at time zero; CFₜ is the net cash flow in period t; r is the discount rate per period; and T is the evaluation horizon. Cash flows and the discount rate must use the same currency and time units. NPV assumes that cash-flow estimates and the discount rate adequately represent risk and that intermediate cash flows can be reinvested at approximately the discount rate.

A positive NPV indicates value creation relative to the required return, subject to the quality of the forecasts and the chosen discount rate. NPV is generally preferable to ranking projects by percentage return when projects differ in scale.

Profitability index

For an independent project with positive initial investment, the profitability index is:

A PI above 1 indicates positive NPV under the same assumptions. PI can be useful when capital is rationed, but ranking by PI alone can favor small projects and produce a suboptimal portfolio when projects are indivisible or interdependent.

Payback period

Payback is the time required for cumulative cash inflows to recover the initial investment. It is easy to communicate and useful for liquidity concerns, but it ignores cash flows after payback and, in its undiscounted form, ignores the time value of money. Discounted payback addresses the latter but still ignores subsequent value.

Expected monetary value

When outcomes are discrete and probabilities are available, expected monetary value is:

where pₛ is the probability of scenario s, Vₛ is the value in that scenario, and the probabilities sum to 1. EMV is a probability-weighted average, not a prediction that the average outcome will occur. It is most informative when probabilities are credible and decision-makers are able to bear the distribution of outcomes.

Worked financial example

An initiative requires an investment of $1.0 million today and is expected to produce net cash flows of $0.45 million in year 1, $0.50 million in year 2, and $0.40 million in year 3. At a discount rate of 10%:

The present values are approximately $0.409 million, $0.413 million, and $0.301 million. Therefore:

The estimated NPV is approximately $123,000. This result is attractive only relative to the assumptions: forecast volume, margin, launch timing, cannibalization, working capital, taxes, and the selected discount rate.

4. Uncertainty, Scenarios, and Risk

Forecast uncertainty

Innovation forecasts are often uncertain because the organization lacks reliable information about adoption, price, competitive response, technical performance, or regulation. Precision in a spreadsheet does not eliminate uncertainty. Managers should distinguish:

• risk: outcomes are uncertain but probability distributions are reasonably estimable;
• ambiguity: the relevant probabilities or even the outcome set are unclear;
• unknowns: important factors have not yet been identified.

Scenario analysis

Scenario analysis evaluates an initiative under coherent combinations of assumptions. A useful set might include downside, base, and upside scenarios. Each scenario should specify variables such as adoption, price, costs, launch delay, competitor response, and regulatory conditions.

Scenario analysis is not the same as changing one variable at a time. A recession scenario may simultaneously reduce demand, increase financing costs, and delay implementation. Scenarios should be internally consistent and strategically plausible.

Sensitivity analysis

Sensitivity analysis changes one or more assumptions to determine which variables most affect value. For a project with NPV dependent on demand, price, cost, and timing, sensitivity analysis may reveal that launch delay is more consequential than a small change in unit cost. This guides research and mitigation.

Risk-adjusted evaluation

Risk can be incorporated through probability-weighted cash flows, a risk-adjusted discount rate, certainty equivalents, or explicit downside costs. These methods are not automatically equivalent. Increasing the discount rate is a blunt treatment that may undervalue long-term strategic investments and does not represent operational flexibility well. Double-counting risk—by reducing cash flows and increasing the discount rate for the same uncertainty—must be avoided.

A portfolio view also matters. Two initiatives with identical standalone risk may have different portfolio effects if their outcomes are positively correlated. A portfolio concentrated in one market, technology, supplier, or customer segment is more fragile than a diversified portfolio, even if its average forecast is higher.

5. Real Options in Product and Innovation Portfolios

Traditional NPV often assumes a passive investment: commit resources now and receive forecast cash flows later. Innovation decisions are frequently active. Managers can delay, stage, expand, abandon, switch, or pause an initiative as information arrives. This flexibility has option value.

A real option is the value of managerial flexibility concerning a real asset or initiative. Examples include:

• funding a prototype before committing to full commercialization;
• preserving the right to enter a market after regulatory clarification;
• designing a platform that permits later product expansion;
• abandoning a failing pilot before incurring scale-up costs;
• reserving scarce capacity while demand evidence is gathered.

A staged investment can be represented conceptually as:

This is a conceptual relationship rather than a universal valuation formula. Estimating option value requires assumptions about uncertainty, decision timing, exercise costs, and the information that will become available.

Real-options reasoning is especially useful when uncertainty is high, downside can be limited, and learning is material. It is less valuable when investment is irreversible, information will not improve, or delay causes rapid loss of market position.

A common managerial error is to call every project an option. A cheap experiment is not automatically valuable: it must generate decision-relevant information or preserve a credible future choice. Experiments should therefore define learning objectives, evidence thresholds, and stop/continue criteria before launch.

6. Constrained Portfolio Optimization

Why individual ranking fails

Suppose three initiatives are ranked by NPV. Selecting the top two may exceed the engineering team’s capacity, depend on the same platform, or create excessive exposure to one customer segment. Portfolio selection requires a feasible combination.

A simplified binary optimization model is:

Here, vᵢ is the value of initiative i; xᵢ equals 1 if the initiative is selected and 0 otherwise; cᵢ is required budget; B is available budget; hᵢ is required capacity, such as engineering hours; and H is available capacity. The model assumes additive values and divisible resource totals but indivisible project decisions. It is a binary knapsack-like formulation.

Dependencies can be added. If initiative j requires platform initiative i:

Mutually exclusive alternatives can be represented by:

If a minimum strategic commitment requires at least one initiative from a group G:

These formulations are simplified. Real organizations may need integer variables for team allocations, continuous variables for capacity, nonlinear demand effects, timing decisions, scenario constraints, and multi-period budgets.

Illustrative allocation problem

A firm has a $10 million budget and 12,000 specialist hours. Four candidate initiatives have the following estimates:

Selecting A and B yields value 14, uses the full $9 million budget, and requires 10,000 hours. Selecting A and D yields value 12, uses $7 million, and requires 8,000 hours. Selecting B, C, and D yields value 15, uses $9 million, and requires 11,000 hours. Thus, under the stated additive estimates and constraints, B + C + D is superior to A + B, despite A being the highest-value individual initiative. If C is a prerequisite for B, the combination is also dependency-feasible. If C and B compete for the same team in a way not captured by total hours, the model must be refined.

The example illustrates why portfolio optimization is not merely a ranking exercise. It also illustrates model risk: the answer changes if values are interdependent, if capacity is skill-specific, or if the platform produces benefits beyond the three listed projects.

Capacity is multidimensional

Aggregate hours can conceal bottlenecks. Relevant capacity categories may include:

• machine-learning specialists;
• regulatory and quality staff;
• brand and design teams;
• sales enablement capacity;
• manufacturing line time;
• customer-service support;
• executive attention.

A more realistic model uses a constraint for each resource type. It may also use time periods because a team available in year 2 cannot necessarily solve a year-1 bottleneck.

7. Managing Dependencies and Strategic Trade-offs

Dependencies should be documented in a portfolio dependency map. Common types include:

• technical: one product requires a shared architecture or API;
• commercial: a channel launch depends on partner contracting;
• sequencing: research must precede development;
• capacity: two projects compete for the same scarce expertise;
• market: one offering changes demand for another;
• brand: inconsistent positioning creates dilution or confusion.

Dependency management requires more than recording arrows. For each dependency, managers should identify the owner, required date, failure consequence, mitigation, and decision authority.

Strategic trade-offs commonly include:

1. Growth versus reliability. Expansion may increase revenue while reducing service quality.
2. Efficiency versus experimentation. Utilization targets can suppress learning investments.
3. Cannibalization versus defense. A new product may reduce current sales but prevent competitors from capturing customers.
4. Global scale versus local relevance. Standardization lowers cost, whereas adaptation may increase adoption.
5. Speed versus control. Rapid launch may generate learning but increase compliance and brand risk.

A robust portfolio process makes these trade-offs visible and records why an apparently inferior initiative is retained—for example, because it protects a strategic option or reduces existential risk.

8. Governance, Decision Rights, and Ethical Control

Portfolio governance should define who proposes, evaluates, funds, pauses, and terminates initiatives. A stage-gate system may include discovery, concept validation, business case, development, pilot, launch, and post-launch review. Gates should be evidence-based, not ceremonial.

Useful governance practices include:

• separating advocacy from independent review;
• publishing criteria and decision logs;
• requiring assumptions and confidence ranges;
• reviewing forecast accuracy after launch;
• reallocating funds as evidence changes;
• establishing explicit kill criteria;
• protecting dissenting expert views;
• using “sunset” reviews for mature products.

Ethical issues include discriminatory targeting, privacy-invasive experimentation, unsafe products, deceptive sustainability claims, and unequal access created by allocation choices. A high financial return does not justify violating legal, safety, privacy, or human-rights obligations. Compliance should be treated as a threshold condition where appropriate, not simply as a low-weight criterion that can be offset by commercial attractiveness.

AI can support portfolio work by extracting evidence, forecasting demand, identifying dependencies, simulating scenarios, and detecting anomalies in project performance. However, AI-generated scores may reproduce historical bias, confuse correlation with causation, or conceal uncertain assumptions. Human accountability remains essential. Models should be documented, validated against out-of-sample evidence where possible, monitored for drift, and reviewed for fairness and explainability.

9. Detailed Illustrative Case: Aurora Home Systems

Background

Aurora Home Systems, a fictional smart-home brand, sells connected thermostats, security devices, and energy-monitoring services. Its product portfolio is profitable but fragmented. The company has $12 million available for the next planning cycle and limited cybersecurity, firmware, and service-design capacity.

Management is considering four initiatives:

• A: Premium thermostat redesign, strengthening the core franchise;
• B: Low-cost market-entry device, targeting price-sensitive households;
• C: Home-energy analytics platform, enabling subscription services;
• D: Cybersecurity and privacy architecture, reducing exposure and supporting future products.

Problem

A preliminary ranking favors A and B because their revenue forecasts are attractive. However, B requires the platform created by C, while A and C compete for the same firmware specialists. D has modest direct revenue but is necessary for compliance with a forthcoming privacy standard and reduces the probability and impact of a major incident.

Analysis

The team first applies mandatory gates. Any initiative must meet privacy and security requirements; therefore, D is not scored merely as an optional revenue project. The team then constructs scenarios for adoption and launch timing. Under the downside scenario, B performs poorly because price competition is intense, whereas C retains value through enterprise partnerships. The team also identifies that C has option value: a successful pilot could support energy-management services, utility partnerships, and future devices.

A weighted score is used for the discretionary comparison, with criteria for strategic fit, expected economic contribution, customer value, feasibility, and learning value. Scores are reviewed by finance, engineering, legal, marketing, and customer operations rather than by the sponsoring teams alone.

The optimization model includes budget, specialist hours, and the dependency constraint that B cannot be selected without C. It also includes a minimum cybersecurity requirement represented by D. The recommended portfolio is D + C + A, with B deferred until C’s pilot reaches predefined adoption and reliability thresholds.

Approach and solution

Aurora funds D as a risk and capability investment, launches a limited C pilot with utilities and high-engagement customers, and redesigns A only after resolving the shared firmware bottleneck. B is preserved as a staged option rather than fully funded. Its next decision depends on pilot evidence, unit economics, security performance, and channel commitments.

Outcome

The portfolio sacrifices the highest near-term unit-volume forecast but improves strategic coherence and resilience. It reduces dependency risk, creates a subscription option, and preserves the possibility of entering the price-sensitive segment without committing full scale-up capital prematurely.

Learning points

1. A project’s standalone ranking can be misleading when dependencies are material.
2. Risk-reduction investments may be strategically mandatory even without high direct NPV.
3. Staging converts some uncertainty into information before irreversible spending.
4. Portfolio value includes capability, resilience, and option creation.
5. Governance quality is part of portfolio performance, not an administrative afterthought.

10. Best Practices and Limitations

Effective portfolio management combines quantitative discipline with strategic judgment. Recommended practices are:

• begin with explicit strategic choices and non-negotiable constraints;
• distinguish mandatory investments from discretionary growth bets;
• use multiple lenses: strategic score, financial value, risk, and option value;
• model dependencies, cannibalization, and shared resources;
• test rankings under alternative assumptions;
• use ranges and scenarios rather than single-point forecasts;
• allocate exploration funding in stages;
• establish evidence thresholds before experiments begin;
• monitor realized outcomes and update assumptions;
• maintain a visible record of trade-offs and decision rights.

Limitations remain. Models may omit cultural resistance, competitor adaptation, political influence, and emergent opportunities. Expected values can obscure catastrophic downside, while scoring systems can create an illusion of objectivity. Optimization software may produce mathematically feasible but strategically absurd solutions if objectives and constraints are poorly specified. Finally, over-frequent reprioritization can create instability and destroy the very capacity required to execute the portfolio.

Chapter Summary

Portfolio optimization selects combinations of products and initiatives rather than simply ranking individual projects. Weighted scoring models create transparent comparisons, but their validity depends on well-defined criteria, consistent scales, credible evidence, and sensitivity analysis. Financial tools such as NPV, profitability index, payback, and expected monetary value answer different questions and must be applied with consistent time, currency, and risk assumptions.

Scenario and sensitivity analysis expose uncertainty and identify the assumptions most worth investigating. Real-options reasoning recognizes that staged investment, abandonment, delay, and expansion can create value when information improves over time. Constrained optimization formalizes budget, capacity, dependency, and exclusivity relationships, while managerial interpretation remains necessary because portfolio value is often interdependent and difficult to quantify fully.

Strong governance integrates strategic, financial, operational, ethical, and regulatory considerations. It defines decision rights, uses evidence-based gates, records assumptions, protects dissent, and reallocates resources as learning occurs. The best portfolio is not necessarily the one with the largest forecast return; it is the one that creates strategically appropriate value while remaining feasible, resilient, and accountable.

Key Terms

Portfolio optimization: Selection and coordination of initiatives to maximize strategic value under constraints.

Resource allocation: Distribution of scarce money, people, time, technology, and capacity across initiatives.

Weighted scoring model: A structured method that aggregates criterion ratings using explicit weights.

Net present value (NPV): Present value of expected future cash flows minus initial investment.

Profitability index: Present value of future inflows divided by the initial investment.

Payback period: Time required to recover an initial investment from cash inflows.

Expected monetary value: Probability-weighted average of values across defined outcomes.

Scenario analysis: Evaluation of outcomes under internally consistent sets of assumptions.

Sensitivity analysis: Examination of how results change when assumptions vary.

Real option: Value arising from managerial flexibility over a real investment.

Stage-gate: A governance process in which initiatives pass evidence-based decision points.

Dependency: A relationship in which one initiative relies on another initiative, resource, or condition.

Cannibalization: Reduction in sales of an existing offering caused by a new offering.

Capacity constraint: A limit on available resources required for execution.

Strategic fit: Degree to which an initiative supports organizational strategy and positioning.

Risk: Uncertainty associated with outcomes whose probability distribution is sufficiently estimable.

Ambiguity: Uncertainty in which probabilities, outcomes, or causal relationships are unclear.

Option value: Value of preserving the ability to make advantageous future decisions.

Portfolio diversification: Distribution of initiatives across markets, technologies, or risk factors to reduce concentration.

Governance: Decision rights, controls, processes, and accountability used to manage the portfolio.

Review Questions

Short Answer Questions

1. Why can the highest-NPV initiatives fail to form the best portfolio?
2. What assumptions underlie an additive weighted scoring model?
3. Distinguish risk from ambiguity in portfolio decision-making.
4. Give two examples of real options in product management.
5. Why should dependencies be represented explicitly in an optimization model?

Long Answer Questions

1. Explain how strategic scoring and financial evaluation should be integrated without allowing either to dominate automatically.
2. Compare NPV, profitability index, payback period, and expected monetary value, including the managerial question each metric addresses.
3. Discuss how scenario analysis, sensitivity analysis, and staged investment can improve innovation decisions.
4. Explain the limitations of binary portfolio optimization models and propose extensions for real organizational settings.
5. Analyze the ethical and governance risks associated with using AI to prioritize product initiatives.

Analytical / Application Questions

1. A company has a fixed budget, three specialist capacity constraints, and six initiatives with technical dependencies. Design the variables, objective, and at least five constraints for a suitable portfolio model.
2. Construct a three-scenario evaluation for a proposed subscription product. Identify the assumptions, probabilities if appropriate, key sensitivity variables, and a decision threshold for proceeding.
3. Using the Aurora Home Systems case, defend an alternative portfolio to D + C + A. State the strategic rationale, constraints, risks, and evidence that would cause you to revise the recommendation.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes portfolio optimization from project prioritization?

A. Portfolio optimization selects the initiative with the highest individual NPV.

B. Portfolio optimization selects a feasible combination while considering interactions and constraints.

C. Portfolio optimization excludes qualitative strategic criteria.

D. Portfolio optimization applies only to mature products.

Correct Answer: B. Portfolio optimization selects a feasible combination while considering interactions and constraints.

Question 2

In a weighted scoring model, what condition is normally imposed on criterion weights?

A. All weights must be greater than 1.

B. Weights must be negative for cost criteria.

C. Weights normally sum to 1 and are non-negative.

D. Weights must all be identical.

Correct Answer: C. Weights normally sum to 1 and are non-negative.

Question 3

An initiative requires $1.0 million initially and produces cash flows of $0.45 million, $0.50 million, and $0.40 million in years 1–3. At 10%, its approximate NPV is:

A. -$0.123 million

B. $0.123 million

C. $0.877 million

D. $1.123 million

Correct Answer: B. $0.123 million

Question 4

Which feature most strongly indicates that real-options reasoning is useful?

A. The investment is fully reversible and produces no new information.

B. Managers can stage investment and learn before making an irreversible commitment.

C. All future cash flows are certain.

D. The project has no decision points after approval.

Correct Answer: B. Managers can stage investment and learn before making an irreversible commitment.

Question 5

In the constraint x_j ≤ x_i, what relationship is represented?

A. Initiative i can be selected only if j is selected.

B. Initiatives i and j must both be rejected.

C. Initiative j can be selected only if i is selected.

D. Initiatives i and j must receive equal funding.

Correct Answer: C. Initiative j can be selected only if i is selected.

Question 6

What is a principal limitation of the undiscounted payback period?

A. It requires probability estimates for every cash flow.

B. It ignores cash flows after the recovery point and the time value of money.

C. It cannot be calculated for projects with initial investment.

D. It always produces the same ranking as NPV.

Correct Answer: B. It ignores cash flows after the recovery point and the time value of money.

Question 7

Why might a cybersecurity initiative be treated as a mandatory gate rather than a weighted scoring criterion?

A. Cybersecurity never affects financial value.

B. Legal or safety requirements may be non-negotiable and cannot be offset by other benefits.

C. Cybersecurity initiatives always have the highest NPV.

D. Gates eliminate the need for governance.

Correct Answer: B. Legal or safety requirements may be non-negotiable and cannot be offset by other benefits.

Question 8

What does expected monetary value represent?

A. The guaranteed outcome of a project.

B. The worst possible outcome under all scenarios.

C. A probability-weighted average of values across defined outcomes.

D. The discount rate used in NPV analysis.

Correct Answer: C. A probability-weighted average of values across defined outcomes.

Question 9

Which situation illustrates a portfolio dependency?

A. Two unrelated products have different brand names.

B. A new device requires a shared platform funded by another initiative.

C. A product has a positive gross margin.

D. A mature product enters a declining market.

Correct Answer: B. A new device requires a shared platform funded by another initiative.

Question 10

Why can a mathematically optimal model output still require managerial review?

A. Optimization models never use constraints.

B. The model may omit strategic interactions, ethical concerns, or poorly specified assumptions.

C. Mathematical models cannot represent budgets.

D. Managers are legally prohibited from using optimization.

Correct Answer: B. The model may omit strategic interactions, ethical concerns, or poorly specified assumptions.