Group Visits

What Group Visits measures

Group Visits is a weighted preference that counts how often selected people are assigned to selected groups across selected sessions. It is useful when group identity matters: activity stations, departments, hosts, rooms, or destinations are not interchangeable.

A valid schedule may still miss a Group Visits target. The solver includes the resulting weighted penalty in its optimization cost; it does not treat the target as a hard requirement.

Choose the scope

The editor provides three scopes:

• People — selected participants, or all people when the scope is left as all.
• Groups — selected destinations, or all groups.
• Sessions — selected periods, or all sessions.

Exclude stationary facilitators from a rotating-participant target. If a facilitator is selected and fixed to one station, an all-people exact-once target also asks that facilitator to rotate.

Availability still matters. A participant can accumulate visits only in selected sessions they attend.

Choose how groups are counted

With separate counting, selecting four people and three groups creates twelve person-and-group measurements. With combined counting, it creates four measurements—one per person.

A combined count increases by at most one in a session because a person has one assignment. It answers whether the assigned group belongs to the selected set, not how many selected groups exist in that session.

Choose visit bounds

The Visit target control supports:

• Exactly a visit count — lower and upper are the same.
• Between lower and upper visits — counts inside the range have no raw deviation.
• At least a lower visit count — only a shortfall is penalized.
• At most an upper visit count — only excess visits are penalized.

Counts are non-negative whole numbers. An exact or lower count cannot exceed the number of selected sessions. That form validation catches a local impossibility, but it cannot prove the entire scenario is feasible.

Check feasibility before solving

For separate counting, if each selected person must visit each of G selected groups exactly V times, that person needs at least G × V attended, selected sessions because one person can visit only one group per session.

Capacity must also cover the requested visits. Across the selected sessions, a station needs enough available places to receive every targeted participant the requested number of times.

Be especially careful when the selected groups are all available groups. If a person attends more sessions than the total allowed visits across those groups, every extra assignment must repeat a selected group and exceed an upper bound.

Requirements can further reduce feasibility. Fixed assignments, together/apart rules, availability, and group capacities all apply while the solver tries to reduce the Group Visits penalty.

Pure SGP solver families reject Group Visits because those solvers treat groups as unlabeled contact geometry. Use a scenario-capable solver path that supports named destination constraints; backend validation remains the source of truth for solver-family support.

Choose weight and penalty function

Group Visits calculates a raw distance outside the configured bounds. If a count is below the lower bound, the distance is the shortfall. If it is above the upper bound, the distance is the excess.

• Linear cost for one person/group measurement is weight × distance.
• Squared cost for one person/group measurement is weight × distance².

GroupMixer applies that calculation to each person-and-group measurement in separate mode, or each person measurement in combined mode, then sums the costs into the constraint penalty.

Squared scoring distinguishes one large miss from several small misses more strongly. Weight controls importance relative to other preferences and objectives. Neither setting makes Group Visits mandatory.

Worked example: four stations

Twenty-four students rotate through four named stations over four periods. Each station has capacity six. The target applies to all students, all four stations, and all four periods, with Each selected group separately and Exactly 1 visit.

The requested visit count is:

24 students × 4 stations = 96 student-to-station visits

The schedule provides:

4 periods × 4 stations × 6 places = 96 assignments

The counts align exactly, so the visit target is feasible before considering other constraints. A result with no Group Visits deviations assigns every student to every station once. It may still contain repeated classmates because destination coverage and contact variety are different goals.

Use the explicit example action on this topic to create the same scenario as a separate saved workspace. It opens Group Visits for inspection and does not overwrite the current scenario.

Read the compliance card

The Group Visits compliance card reports raw evidence for each miss:

• participant;
• station or selected station set;
• actual visit count;
• requested lower and upper bounds;
• configured session scope;
• raw bound deviation.

The card does not attribute a missing visit to one uniquely causal session, and it does not expand weight and linear-versus-squared cost for every deviation. “No violations” means the configured visit bounds were met. It does not prove that transitions are practical or that each activity received the right human mix.

Replan after periods have happened

A new solve does not automatically import completed visit history. If the event is already underway, do not simply apply the original exact-once target to the remaining periods; that can send people back to stations they already visited.

Either regenerate the full schedule and communicate every changed period, or create a remaining-period scenario with Group Visits targets that describe each participant’s outstanding destinations. Different visit histories may require separate people scopes and targets.

For a complete operational workflow, read Run station rotations for groups. For score and compliance interpretation, continue to Understand results.