[Canonical page]() [Full documentation index]() # Evaluating alternatives This guide shows how to compare several possible edits in one call and pick the best one. For example, you can check which resource should take a task. It covers building candidates, scoring them with and without leveling, reading the metrics, handling candidates that fail, and committing the winner. ## How evaluation works [`project.evaluate`]() takes an array of **candidates**. A candidate is an object with an `id` and a `changes` field. `changes` is a [`SchedulingChanges`]() object, the same change set you'd pass to [`apply`](). Each candidate is one alternative: "give task B to dev2", "split the review between two people", "add a link". The call scores every candidate against the project's accepted baseline and returns the scores. The project isn't modified, just like with [`preview`](). Every candidate goes through the same run on a disposable copy of the project: 1. Copy the reference state. That's the accepted baseline, or its cleared copy with `clearFirst` (see [Options]()). 2. `apply` the candidate's changes. 3. [`solve(null)`]() and measure the **scheduled** stage. 4. When the `level` option is set, [`level`]() with those options and measure the **leveled** stage. This is the same sequence you'd run to commit the candidate for real. Each candidate starts from the reference state, never from the previous candidate. Both stages are compared with the one baseline. The result has three parts: - `baseline`: the metrics of the reference state itself. All its deltas are 0. - `candidates`: one entry per candidate, in input order. Each entry holds `scheduled` metrics, `leveled` metrics (only with `level`) and a `cost`. - `baselineId`: the token of the accepted baseline the candidates were scored against. Each stage is described by a [`Metrics`]() object: the project finish and the finish of each focus task, resource overload, conflicts, cycles and a single `cost` number. Every measurement is an absolute `value` plus a `delta` against the baseline, so you can show either one. Errors come in two levels. Problems with the call itself throw and fail the whole call: - no solved baseline: [`ProjectError`]() with code `"no-baseline"` - a stale `baselineId` in the options or in any candidate: `ProjectError` with code `"stale-baseline"` - an unknown focus task, or leveling options that don't fit the baseline: [`InputError`]() Everything else fails only its own candidate. That candidate gets an `error` and no metrics, and the other candidates are scored as usual. `evaluate` reports scores, not dates. The result has no task updates and no overallocation windows. Once you've picked a candidate, commit it yourself with `apply` and `level`, or look at its dates with `preview`. ## Comparing candidates All examples on this page use the same small project: tasks A and B, two working days each, both starting on Monday, August 10, 2026, and both assigned to `dev1`. The resource `dev2` is idle. `dev1` is overloaded for two days. ### Candidates Build each alternative as a `{ id, changes }` object and pass them all to [`evaluate`](). An empty change set is a useful candidate too: it scores "leave the plan as it is" next to the other options. ```js import { Project } from "plen-js"; const project = new Project(); project.solve({ tasks: [ { id: "A", start_date: "2026-08-10 08:00", duration: 2 }, { id: "B", start_date: "2026-08-10 08:00", duration: 2 }, ], resources: [{ id: "dev1" }, { id: "dev2" }], assignments: [ { id: "aA", task: "A", resource: "dev1", value: 1 }, { id: "aB", task: "B", resource: "dev1", value: 1 }, ], }); const result = project.evaluate( [ // Candidate 1: keep the plan as it is. { id: "stay", changes: {} }, // Candidate 2: give task B to dev2. { id: "dev2", changes: { assignments: { update: [{ id: "aB", resource: "dev2" }] } } }, ], { focus: "B" }, ); console.log(`baseline: cost ${result.baseline.cost}`); for (const candidate of result.candidates) { console.log(`${candidate.id}: cost ${candidate.cost}`); } ``` Output: ``` baseline: cost 2 stay: cost 2 dev2: cost 0 ``` Without leveling, only the scheduled stage is scored. `stay` keeps both tasks on `dev1` and costs 2, two resource-days of overload. `dev2` removes the overload and costs 0. The candidate `id` is echoed back so you can match results to inputs. Ids aren't checked for uniqueness, and results keep input order anyway. See [`project.evaluate`](). ### Options The options object has four fields, all optional: - `level`: a [`LevelingOptions`]() object levels each candidate after scheduling it and adds the leveled stage. `false` (the default) scores the scheduled stage only. `level: true` is invalid and throws an [`InputError`](); pass `{}` for the default leveling options. - `focus`: the tasks to report finish dates and focus overload for. There's no implicit "modified task", so list the tasks your candidates are about. - `baselineId`: a baseline token. The call throws [`ProjectError`]() with code `"stale-baseline"` when it no longer names the accepted baseline. Each candidate's `changes.baselineId` is checked the same way. - `clearFirst`: score against a copy with the in-scope `leveling_delay` and `pauses` cleared. :::note `focus` takes one task id or an array of ids. ::: Leveling changes the picture. Without it, an overloaded plan looks bad on overload. With it, the overload turns into delay, which is what the user gets after committing and leveling: ```js import { Project } from "plen-js"; // Tasks A and B, two days each, both on dev1 in the same week. function solvedProject() { const project = new Project(); project.solve({ tasks: [ { id: "A", start_date: "2026-08-10 08:00", duration: 2 }, { id: "B", start_date: "2026-08-10 08:00", duration: 2 }, ], resources: [{ id: "dev1" }, { id: "dev2" }], assignments: [ { id: "aA", task: "A", resource: "dev1", value: 1 }, { id: "aB", task: "B", resource: "dev1", value: 1 }, ], }); return project; } function reassign(assignment, resource) { return { assignments: { update: [{ id: assignment, resource }] } }; } const project = solvedProject(); const result = project.evaluate( [ { id: "stay", changes: {} }, { id: "dev2", changes: reassign("aB", "dev2") }, ], { level: {}, focus: ["A", "B"] }, ); for (const candidate of result.candidates) { const leveled = candidate.leveled; if (!leveled) continue; const finish = leveled.finish.project; console.log(`${candidate.id}: project ends ${finish.date} (${finish.delta} days), cost ${candidate.cost}`); for (const task of leveled.finish.tasks) { if (task.finish) console.log(` ${task.id} ends ${task.finish.date} (${task.finish.delta})`); } } ``` Output: ``` stay: project ends 2026-08-14 08:00 (2 days), cost 2 A ends 2026-08-12 08:00 (0) B ends 2026-08-14 08:00 (2) dev2: project ends 2026-08-12 08:00 (0 days), cost 0 A ends 2026-08-12 08:00 (0) B ends 2026-08-12 08:00 (0) ``` For `stay`, leveling resolves the overload by pushing B two days later, so the cost is now the delay. For `dev2` there's nothing to level. `level` runs with the options exactly as you pass them. [`LevelingOptions.clearFirst`]() defaults to `true`, so each candidate is re-leveled from zero. After a previous `level` run, this creates a trap. The stored delays are already in the baseline finish, and they hide the overload the current assignments cause. A candidate that removes the overload then shows a negative finish delta, because leveling from zero drops the old delays. The `clearFirst` option of `evaluate` fixes this. It replaces the reference state with a copy that has the in-scope delays and pauses cleared, solved again. The baseline metrics and every candidate run then use that copy. The scope is `level.scope.tasks` when you give it, otherwise every task. It's independent of `LevelingOptions.clearFirst`, so it also works without `level`. ```js import { Project } from "plen-js"; // Tasks A and B, two days each, both on dev1 in the same week. function solvedProject() { const project = new Project(); project.solve({ tasks: [ { id: "A", start_date: "2026-08-10 08:00", duration: 2 }, { id: "B", start_date: "2026-08-10 08:00", duration: 2 }, ], resources: [{ id: "dev1" }, { id: "dev2" }], assignments: [ { id: "aA", task: "A", resource: "dev1", value: 1 }, { id: "aB", task: "B", resource: "dev1", value: 1 }, ], }); return project; } function reassign(assignment, resource) { return { assignments: { update: [{ id: assignment, resource }] } }; } const project = solvedProject(); // The plan was leveled earlier: B carries a two-day leveling delay. const leveled = project.level(); const result = project.evaluate([{ id: "dev2", changes: reassign("aB", "dev2") }], { baselineId: leveled.baselineId, level: {}, clearFirst: true, }); // The cleared baseline shows the overload that B's delay was hiding. console.log(`baseline overload: ${result.baseline.overload.total.value}`); for (const candidate of result.candidates) { const finish = candidate.leveled?.finish.project; if (finish) console.log(`${candidate.id}: finish delta ${finish.delta}`); } ``` Output: ``` baseline overload: 2 dev2: finish delta 0 ``` With `clearFirst`, the baseline reports 2 resource-days of overload and the `dev2` candidate a finish delta of 0. Without it, the baseline has no overload and the same candidate shows a finish delta of -2. Both views are valid. `clearFirst` is the one that matches a commit with `level` and `clearFirst`, which re-levels from zero. :::note The leveling result is a [`ScheduleResult`]() with an added `rounds` field; read `updates` on it. ::: A few details: - `baselineId`, candidate tokens, `level` options and focus ids are all checked against the accepted state before anything is cleared. `clearFirst` adds no new errors. - A candidate's own `changes.baselineId` is dropped before its trial run, because the cleared copy has a different token. The result's `baselineId` is still the accepted state's. - When there's nothing to clear in scope, the output is the same as with `clearFirst: false`. The full option list is on the [`project.evaluate` reference](). For the leveling options themselves, see [Leveling resources](). ### Metrics and cost Every stage reports the same [`Metrics`](): | Field | What it measures | |---|---| | `finish.project` | Latest end over scheduled tasks. `null` when the stage has no tasks. | | `finish.tasks` | One `{ id, finish }` entry per focus task, in focus order. | | `overload.total` | Overload over all resources. | | `overload.resources` | Overload per resource, for every resource that's non-zero in the stage or the baseline. | | `overload.focus` | Per focus task: the overload per resource in the buckets that intersect the task. | | `conflicts`, `cycles` | How many conflicts and cycles the stage has. | | `cost` | One number for ranking, lower is better. | Dates come as a `Shift`: `date` plus `delta`. The delta is the working time from the baseline finish in `config.duration_unit`, positive when later. A task's delta is measured on its own calendar, the project's on the global calendar. The `date` has the same form as other result dates: a string by default, a `Date` with `new Project({ dates: "date" })` (see [Plain objects and dates]()). Counts and overload come as an `Amount`: `value` plus `delta`. Overload is measured the same way the leveler sees it. For each time bucket, it's how far a resource's load goes over its capacity, summed over buckets. The unit is capacity × buckets, so on a day basis 2 means "two resource-days over". One bucket size (basis) applies to the baseline and every stage of every candidate: `level.basis` when you give it, otherwise the one derived from `config.duration_unit`. The same per-resource sum is what [`loadProfile`]() shows as excess at that basis. Focus overload shows where the trouble is. At the scheduled stage it's the overload around the task. At the leveled stage it's what leveling couldn't resolve there. A milestone has zero length and covers no bucket. The leveled stage has every `Metrics` field plus `converged`, `rounds` and `windows`, the number of overallocation windows left after leveling. :::note Leveled metrics have the same fields as `Metrics` plus `converged`, `rounds` and `windows`. ::: `cost` is one fixed heuristic: ``` cost = max(0, finish.project.delta) + overload.total.value + 1000 × (conflicts.value + cycles.value) ``` When the basis follows `duration_unit`, a day of delay and a resource-day of overload weigh the same. An earlier finish doesn't earn a bonus. The 1000 penalty keeps any broken schedule, with conflicts or cycles, behind every valid one. A candidate's `cost` is the cost of its last stage: leveled when you level, scheduled otherwise. The baseline carries its own `cost`. If the heuristic doesn't fit your product, rank on the raw values instead. This function prints the per-resource overload change of each candidate's scheduled stage: ```js function printOverload(result) { for (const candidate of result.candidates) { if (!candidate.scheduled) continue; for (const row of candidate.scheduled.overload.resources) { console.log( `${candidate.id}: ${row.resource} overload ${row.overload.value} (${row.overload.delta} vs baseline)`, ); } } } ``` For the project above, the `dev2` candidate prints a row for `dev1` with a delta of -2. The row is there because `dev1` was overloaded in the baseline, even though it's zero in the stage. Resources come in stage order, then resources that exist only in the baseline. See the [`Metrics` reference](). ### Failed and cyclic candidates A candidate can fail without failing the call. Three cases need handling: - **Rejected.** `apply` rejects an operation in the change set (the result reports only the first rejection), or `solve` or `level` fails on the candidate's input, for example a leveling scope id the candidate removed. The candidate result has an `error` with a `path` and a `message`. `scheduled`, `leveled` and `cost` are absent. - **Cyclic.** The candidate creates a dependency cycle. `scheduled` is measured on the unchanged dates the cyclic solve returns, and `cycles.value` is above 0. Leveling stops at round 0 on a cyclic plan, so `leveled` is absent. The cycle penalty pushes its cost over 1000. - **Focus task removed.** When the candidate removes a focus task, that task's `finish` and its focus overload `resources` are `null`. The rest of the metrics are normal. ```js import { Project } from "plen-js"; // Tasks A and B, two days each, both on dev1 in the same week. function solvedProject() { const project = new Project(); project.solve({ tasks: [ { id: "A", start_date: "2026-08-10 08:00", duration: 2 }, { id: "B", start_date: "2026-08-10 08:00", duration: 2 }, ], resources: [{ id: "dev1" }, { id: "dev2" }], assignments: [ { id: "aA", task: "A", resource: "dev1", value: 1 }, { id: "aB", task: "B", resource: "dev1", value: 1 }, ], }); return project; } function reassign(assignment, resource) { return { assignments: { update: [{ id: assignment, resource }] } }; } const project = solvedProject(); const result = project.evaluate( [ { id: "missing", changes: reassign("nope", "dev2") }, { id: "cycle", changes: { links: { add: [ { id: "ab", source: "A", target: "B", type: "fs" }, { id: "ba", source: "B", target: "A", type: "fs" }, ], }, }, }, { id: "removed", changes: { tasks: { remove: ["B"] } } }, ], { level: {}, focus: "B" }, ); for (const candidate of result.candidates) { if (candidate.error) { console.log(`${candidate.id} rejected at ${candidate.error.path}: ${candidate.error.message}`); continue; } const scheduled = candidate.scheduled; if (scheduled.cycles.value > 0) { console.log(`${candidate.id} is cyclic, leveled: ${candidate.leveled !== undefined}`); } for (const task of scheduled.finish.tasks) { if (task.finish === null) console.log(`${candidate.id} removes focus task ${task.id}`); } console.log(`${candidate.id} cost ${candidate.cost}`); } ``` Output: ``` missing rejected at assignments.update[0]: unknown assignment id cycle is cyclic, leveled: false cycle cost 1002 removed removes focus task B removed cost 0 ``` Rank only candidates that have a `cost`. The cyclic candidate has one, but the penalty keeps it at the bottom. Its 1002 is the cycle penalty plus the two resource-days of overload `dev1` still has. Call-level errors still throw. An unknown focus id fails the whole call with an [`InputError`]() at path `focus[i]`, and a stale token with a [`ProjectError`]() with code `"stale-baseline"`: ```js import { InputError, Project, ProjectError } from "plen-js"; const project = new Project(); project.solve({ tasks: [{ id: "A", start_date: "2026-08-10 08:00", duration: 2 }] }); try { project.evaluate([], { focus: "X" }); } catch (error) { if (error instanceof InputError) console.log(`${error.path}: ${error.message}`); else if (error instanceof ProjectError) console.log(error.code); else throw error; } ``` Output: ``` focus[0]: unknown task "X" ``` See [Error handling](). ### Sequential only :::note Candidates are scored one after another. ::: Each candidate runs on its own copy, so candidates are independent, but `evaluate` runs them one after another on the calling thread. Like every `Project` method, it's synchronous and CPU-bound, so a large evaluation blocks the event loop until it finishes. In a browser UI or a busy server, move the project to a Worker and call `evaluate` there. See [Run in a worker](). ### Committing the winner [`evaluate`]() never changes the project. To commit the chosen candidate, run its changes for real: [`apply`]() the same change set, then [`level`]() with the same [`LevelingOptions`](), `step` included. With other options, the committed plan isn't the one that was scored. If you scored without `level`, follow `apply` with [`solve(null)`]() instead. ```js import { Project } from "plen-js"; // Tasks A and B, two days each, both on dev1 in the same week. function solvedProject() { const project = new Project(); project.solve({ tasks: [ { id: "A", start_date: "2026-08-10 08:00", duration: 2 }, { id: "B", start_date: "2026-08-10 08:00", duration: 2 }, ], resources: [{ id: "dev1" }, { id: "dev2" }], assignments: [ { id: "aA", task: "A", resource: "dev1", value: 1 }, { id: "aB", task: "B", resource: "dev1", value: 1 }, ], }); return project; } function reassign(assignment, resource) { return { assignments: { update: [{ id: assignment, resource }] } }; } const project = solvedProject(); const candidates = [ { id: "stay", changes: {} }, { id: "dev2", changes: reassign("aB", "dev2") }, ]; const leveling = {}; const result = project.evaluate(candidates, { level: leveling }); // Lowest cost wins; rejected candidates have no cost. let best; result.candidates.forEach((scored, i) => { if (scored.cost === undefined) return; if (!best || scored.cost < best.cost) best = { cost: scored.cost, candidate: candidates[i] }; }); if (best) { // Throws ProjectError("stale-baseline") if the project changed since the evaluation. const report = project.apply({ ...best.candidate.changes, baselineId: result.baselineId }); if (report.rejected.length === 0) { // The same options that scored the candidate. project.level(leveling); console.log(`committed ${best.candidate.id} (cost ${best.cost})`); } } ``` Output: ``` committed dev2 (cost 0) ``` Results come back in input order, so the index of a scored entry is the index of its change set. Setting `baselineId` to the token from the result makes `apply` refuse the commit when another edit landed in between. In that case, evaluate again. If you scored with the `clearFirst` option of `evaluate`, commit with `level` and [`LevelingOptions.clearFirst`]() (the default) so the commit re-levels from zero, like the scoring did. To handle a rejected `apply`, see [Recover from rejected changes](). To generate candidates automatically instead of writing them by hand, use [`suggestAssignments`](), which scores its steps the same way. See [Suggesting assignments]().