References

The industrial planning glossary.

Forty terms, defined without needless jargon — from critical path to generative scheduling.

B

Baseline: The reference version of a schedule, frozen at a given point in time, against which actual progress is compared. It is used to measure schedule and cost variances and to track approved changes.
Bid schedule: A schedule produced during the bid phase to commit the company to a delivery date. Though built quickly, it locks in the contract's margins.

Capacity planning: Consolidating the demand placed by projects against the available capacity of teams and assets, by discipline and by period. It reveals over-allocations before they cause delays.
Concurrent work: Several teams or companies working at the same time in the same area or on the same system. It requires sequencing and safety rules, especially in industrial and nuclear maintenance.
Cost and schedule overrun: The gap between planned and actual cost or duration. Caught early, it can be corrected; caught late, it has to be absorbed.
Cost estimating: Estimating costs and assigning them to the activities of a schedule (labor, equipment, procurement). It ties the timeline to the budget.
CPM: Critical Path Method, formalized in the late 1950s. It calculates the early and late dates of each activity from durations and logic links.
Critical path: The sequence of activities in which any delay pushes back the project finish date. Calculated from logic links, it ignores resource availability by default.

Data pool: A structured body of a company's project data: schedules, lessons learned, engineering data. It is the raw material for analysis and learning.
DCMA 14-point assessment: A set of 14 schedule construction checks formalized by the US Defense Contract Management Agency. It checks logic, links, constraints, float and the critical path, but not whether durations are realistic.
Deep learning: A family of machine learning methods based on multi-layer neural networks. It excels at combining many factors to recognize patterns.
Demand vs. capacity: A comparison between the work to be done (demand) and the resources available to do it (capacity). Demand above capacity signals over-allocation; demand below capacity, underutilization.

EVM: Earned Value Management. It compares work performed, valued at budget, with work planned and costs incurred, to separate schedule variance from cost variance.

Float: The time an activity can slip without delaying another activity (free float) or the project finish (total float). Negative float signals a commitment that can no longer be met.

Gantt chart: A representation of a schedule as horizontal bars along a timeline. Easy to read, it shows dates and links but not how reliable the durations are.
Generative scenario planning: AI-generated, complete schedule scenarios under real-world constraints, compared side by side to support trade-off decisions.
Generative scheduling: Automatic generation of a logic-driven schedule from documents, a prompt or a history of past projects. The planner reviews and validates the result.

Kanban: A method for visually managing the flow of work in columns (to do, in progress, done). Suited to tracking a team's day-to-day tasks.

Learning loop: The loop through which every completed project enriches the data and improves the next forecasts. It turns history into a compounding advantage.
Lessons learned: Insights drawn from past projects: actual durations, sequences, overruns and their causes.
LLM: Large Language Model, trained on vast bodies of text. Useful for reading and writing, it reasons on generalities rather than on a company's own history.

MCP: Model Context Protocol, an open protocol that gives an enterprise LLM governed access to applications and their data.
Milestone: A zero-duration event that marks a key step: a review, a delivery, a decision. Contractual milestones often carry liquidated damages.
Monte Carlo simulation: A simulation method that randomly draws thousands of plausible values for durations and costs to produce a distribution of possible finish dates. It expresses schedule as a probability rather than a single date.
Multi-constraint: Describes planning that simultaneously accounts for logic links, resources, inventory, contractual milestones and concurrent work. This is the normal situation on large industrial programs.

No-code: An approach that lets users configure, analyze or train without writing code. It makes AI accessible to subject-matter experts.

Pattern mining: Automatically searching data for patterns: recurring sequences, typical successions, frequent causes of delay.
PPM: Project Portfolio Management: selecting and prioritizing projects, allocating resources and tracking value.
Predictability: The ability to announce a date or a cost and to meet it. It is measured by the gap between planned and actual, project after project.
Project controls: The set of disciplines for keeping projects under control: planning, estimating, cost control, risk management and change management.
Prompt: A natural-language instruction given to an AI. In planning, it can describe a need from which a schedule is generated.

Rescheduling: Updating a schedule after a disruption or a change. Fast and scenario-based, it limits the impact; improvised, it just moves the problem elsewhere.
Resource leveling: Shifting activities to remove resource over-allocations. Tools do this with simple priority rules, without searching for the best overall solution.

S-curve: An S-shaped curve showing cumulative planned and actual work or costs over time. A classic EVM visual.
Schedule defect: A flaw that makes a schedule wrong or fragile: a missing link, an unrealistic duration, an unjustified constraint, inconsistent float. It is paid for during execution.
Schedule quality check: Automated schedule quality control that combines standard checks with company-specific rules and suggests corrections.
Sovereign AI: AI whose data, models and hosting are controlled by the user, under a jurisdiction of their choice. It stands in contrast to general-purpose AI services whose operation and hosting are beyond the customer's control.

WBS: Work Breakdown Structure, the hierarchical breakdown of a project's work into work packages. A tool for scope and control.
What-if: Simulating a variant ("what happens if?") to measure its effect on the schedule. Done manually, it tests one variant at a time.
Workload: The work assigned to a resource over a given period. Compared with capacity, it reveals over-allocation and underutilization.

Ask an OFFOLIO expert. A reasoned answer within 48 hours.