Decision Support Systems – concepts, types, architecture, and hospitality applications

1. What is a Decision Support System?

A Decision Support System (DSS) is an interactive computer‑based system that assists managers in solving semi‑structured or unstructured problems by providing timely information, analytical models, and visualisation tools. Unlike routine Transaction Processing Systems (TPS), a DSS does not automate the decision; it augments human judgment.

Key attributes:

Core Components

1. Database Management Subsystem – stores raw data, historical records, and external feeds.
2. Model Management Subsystem – libraries of analytical, statistical, and optimisation models.
3. User Interface Subsystem – dashboards, query forms, and visualisation widgets.
4. Knowledge Base (optional) – rules, heuristics, or expert‑system components for knowledge‑driven DSS.

2. Types of Decision Support Systems

Worked Example: Model‑driven DSS for Room‑Rate Optimisation

A 120‑room boutique hotel wants to maximise RevPAR (Revenue per Available Room) for the upcoming weekend.

1. Input data – historical demand curve, competitor rates, booking lead time, and expected events.
2. Model – a linear programming formulation:

$$\max \sum _{i=1}^n(p_i\times x_i)\text{s.t.}\sum _{i=1}^n{x}_i\le 120,x_i\le {\text{forecast demand}}_i$$

where $p_i$ = price tier $i$, $x_i$ = rooms allocated to tier $i$.

3. Solution – the DSS suggests 30 rooms at $150, 50 rooms at $130, and 40 rooms at $110, projecting a RevPAR of $132.

The manager can instantly adjust assumptions (e.g., a sudden concert) and re‑run the model, observing the impact on RevPAR.

3. DSS Architecture

A typical three‑layer architecture (presentation, logic, data) is illustrated below.

• Presentation Layer: Web portals, mobile apps, or desktop dashboards.
• Logic Layer: Executes models, runs simulations, and applies business rules.
• Data Layer: Centralised data warehouse, often built on a star schema for fast OLAP queries.

4. Decision‑Making Cycle Supported by DSS

The classic Simon’s decision‑making model (Intelligence → Design → Choice → Implementation) maps neatly onto DSS functions.

• Intelligence – DSS pulls data from the PMS, market feeds, and social media to highlight trends (e.g., sudden drop in bookings).
• Design – Users build “what‑if” scenarios using forecasting or optimisation models.
• Choice – The system ranks alternatives based on KPIs (RevPAR, ADR, occupancy).
• Implementation – Recommendations are exported to the PMS or communicated to staff via mobile alerts.

5. Applications in the Hospitality Industry

Real‑World Example: Nabil Bank’s Loan‑Approval DSS

Nabil Bank uses a knowledge‑driven DSS that incorporates credit‑scoring rules, regulatory limits, and market risk models. When a corporate client applies for a loan, the system instantly evaluates the application against the rule base, suggests an interest rate, and flags high‑risk cases for senior manager review. This reduces approval time from days to minutes.

6. Advantages and Disadvantages

7. In the Real World

1. eZee FrontDesk (Hotel PMS) – integrates a data‑driven DSS that visualises occupancy trends and suggests dynamic pricing. The dashboard pulls real‑time booking data and external events (e.g., festivals) to adjust rates on the fly.

2. Daraz Order Fulfilment – uses a model‑driven DSS for warehouse slot allocation. Linear programming decides the optimal picking sequence, reducing order‑to‑delivery time by 15 %.

3. Google Ads Bidding Engine – a knowledge‑driven DSS that applies machine‑learning rules to bid on keywords in real time, maximising ROI for advertisers.

Worked Real Situation: Kathmandu Traffic Routing

The Kathmandu Metropolitan City employs a communication‑driven DSS that aggregates GPS data from NTC’s network, traffic cameras, and citizen reports. The system runs a simulation model to predict congestion hotspots for the next hour and suggests alternate routes to commuters via a mobile app. This reduces average travel time during peak hours by about 12 %.

8. Building a Simple DSS: Step‑by‑Step Trace

Suppose a small boutique hotel wants a basic DSS to decide whether to offer a “late‑checkout” promotion.

1. Define the decision problem – Increase occupancy on low‑demand nights without hurting revenue.
2. Collect data – Past occupancy, average daily rate (ADR), cost of cleaning extra rooms.
3. Choose a model – Simple profit‑impact calculator:

$$\text{Incremental Profit}=(\text{Additional Rooms Sold}\times {\text{ADR}}_{\text{promo}})-(\text{Extra Cleaning Cost})$$

4. Build the interface – A spreadsheet with input cells for “rooms sold with promo” and “promo ADR”.

5. Run scenarios –

   • Scenario A: 5 extra rooms at $80, cleaning cost $10/room → Profit = $5\times 80-5\times 10=\$350$.
   • Scenario B: 8 extra rooms at $75, cleaning cost $12/room → Profit = $8\times 75-8\times 12=\$504$.

6. Choose – Scenario B yields higher profit; manager approves the promotion.

This trace demonstrates the intelligence → design → choice flow in a tangible, low‑tech DSS.

9. Comparison with Other Information Systems

10. Future Trends in DSS for Hospitality

• Artificial Intelligence & Machine Learning – predictive analytics for demand forecasting, sentiment analysis from guest reviews.
• Cloud‑based DSS – scalable, subscription models (e.g., Oracle Hospitality Cloud).
• Mobile‑first interfaces – real‑time alerts to front‑desk staff on smartphones.
• Integration with IoT – sensor data (room temperature, occupancy) feeding into energy‑cost optimisation models.

Exam tip

• Memorise the four DSS types and be ready to match each type to a hospitality example (e.g., data‑driven → occupancy dashboards).
• Diagram the decision‑making cycle and label where the DSS contributes (data collection, model design, choice).
• Comparison tables are a favourite: practice filling the DSS vs. MIS vs. EIS table quickly.
• For scenario questions, follow the structured trace: problem → data → model → alternative analysis → recommendation. Write the steps in bullet form; examiners award marks for clear logical flow.
• Remember the pros & cons list; a two‑column table is easy to recall.

---