Planning concepts in plain language
For planners and operations managers. Each idea in one paragraph, then a worked example.
Items to plan
The things that need a decision. Each has details, such as a name, a duration, a priority, the skills it needs or a deadline. You can add your own details ("needs a forklift licence"): PowerSolver keeps every column of your file.
| Type of plan (step 1) | Items to plan | What PowerSolver decides |
|---|---|---|
| Employee rostering | shift positions | who works each one |
| School timetabling | lessons | the time slot and the room (the teacher is often already chosen) |
| Maintenance scheduling | jobs | the technician and the half-day |
| Job shop scheduling | operations | the machine and the time |
| Task assignment | tasks | who does each task, and when |
| Conference scheduling | sessions | the time slot and the room |
| Cloud optimization | workloads | the server |
| Vehicle routing | visits | which vehicle, and in which order |
| Generic | anything else | 1 to 4 decisions you name |
Decisions
For each item PowerSolver makes one to four choices at once, for example who does an operation, when, in which room and with which anaesthetist. More decisions give more freedom, and a longer search.
Lists of options, and lists of information
A list of options is what a decision chooses from: the staff, the time slots, the rooms. Each option has details the rules use: skills, hours per week, days off, capacity.
A list of information is read, never chosen: the teacher already given to a lesson, the patient of an operation. Rules can still check it, for example "the teacher is not on leave that day".
Rules
Rules come in two strengths, and choosing the right one is your most important job.
| Strength | Meaning | Examples |
|---|---|---|
| Must | never broken: a plan that breaks one is unusable | nobody in two places at once; the right skill; 11 h rest between shifts; a room is big enough |
| Should | a wish: the plan is better when it holds | share the work fairly; keep people's preferred shifts; lowest cost; change as little as possible |
Each rule also has an importance. Between two wishes the more important one wins; Musts always come first.
PowerSolver offers 19 kinds of rules, written as plain sentences in step 4: everything gets a value, no double-booking, skills match, unavailability, capacity, time windows, order of tasks, minimum rest, limits per day or week, days in a row, enough people per time slot, cost, preferences, fair share and more.
Score
Every plan gets a score in two parts, for example 0hard / -12soft. Both count penalties, so closer to 0 is better.
- The first part counts broken Must rules. It must be 0; otherwise the plan cannot be used.
- The second part counts the wishes not met, weighted by their importance.
- To compare two plans, look at the first part, then the second: a plan with fewer broken Musts always wins.
Result
The result is the full list of decisions plus its score, and a verdict in plain words on the Results page. If no Must rule is broken, the plan is valid and can be put into practice.
Two ways to plan
- Assign only (the usual case): each item gets an option. Each shift gets one person, each lesson one room.
- Items in order on a resource: items line up one after another on a person, machine or vehicle, for example visits along a delivery route. PowerSolver then decides the order, and the start and end times follow by themselves.
Choose in step 1. Rule of thumb: if the order of the items on a resource matters, use items in order; otherwise, assign only.
Times that calculate themselves
With items in order, each item's start and end come from the item before it: an item starts when the previous one ends (or when the person's day starts), and ends after its own duration. You never type these times; PowerSolver keeps them up to date as the plan changes, and the Results page uses them.
A worked example
The problem: three development tasks, two developers, three time slots.
| Task | Duration | Skill needed |
|---|---|---|
| Task 1 | 4 h | Java |
| Task 2 | 2 h | React |
| Task 3 | 3 h | Java |
Options: Alice (Java, React) and Bob (Java). Slots: Monday morning, Monday afternoon, Tuesday morning.
Rules:
- Must: the developer has the skill.
- Must: nobody does two tasks at the same time.
- Should (importance 1): tasks in the morning.
The plan: Task 1 → Bob, Monday morning. Task 2 → Alice, Monday morning. Task 3 → Bob, Tuesday morning.
The score: 0hard / 0soft: every Must holds and every task is in a morning. Had Alice been the only Java developer, one task would have moved to the afternoon and the score would be 0hard / -1soft: still usable, one wish not met.
What PowerSolver cannot do
- It chooses from finite lists you give it; it cannot invent options.
- On very large plans it finds a very good plan in the time allowed, not always the mathematically best one.
- It does not set your priorities: you choose the rules, their strength and their importance.
Next steps
- Read PowerSolver in plain words: the steps and how to read results.
- Try a demo: each comes as a ready Excel workbook (Demos).
- Unsure which type of plan you have? Email support@planningpowertools.com.