Task library

31 real tasks, from client brief to delivery.

See what each task asks for, which rules are checked, and how the result is scored. Click a task for its details and diagram.

31 of 31 tasks
  1. TASK 01

    Multi-pot braising schedule

    Food processing Production scheduling Design your own approachSeveral periods or stepsTie-breaking rules

    Meet demand from two order batches without double-booking any pot or time slot. Plans that tie on the main goal are compared on secondary rules.

    What is checked

    Output formatSchedule formatNo resource or time clashesDemand is met

    How it is scored

    The overall score is recalculated with the project’s scheduling goal.

  2. TASK 02

    One-day food production in batches

    Food processing Production scheduling Clear-cut modelSeveral periods or steps

    Fit several production batches into one day within the limits on resources and rounds, while weighing efficiency against order priority.

    What is checked

    Batch schedule formatResource limitsDemand and round limits

    How it is scored

    Scored on production efficiency and priorities.

  3. TASK 03

    Repeated batches of one product

    Food processing Production scheduling Clear-cut modelSeveral periods or steps

    Run batch after batch of a single product in the right process order, replace the braising liquid when it wears out, and finish as early as possible.

    What is checked

    BatchesProcess orderLiquid replaced on timeDemand is met

    How it is scored

    Scored on completion time.

  4. TASK 04

    Five food products over a longer period

    Food processing Production scheduling Design your own approachSeveral periods or stepsTie-breaking rules

    Sequence five products across the pots, where each product depends on what ran before and on the state of the liquid. Finish on time while respecting soft preferences.

    What is checked

    Every product scheduledChangeoversProduct orderTiming rules

    How it is scored

    Scored on completion times, job order and soft preferences.

  5. TASK 05

    Batches linked by shared braising liquid

    Food processing Production scheduling Clear-cut modelSeveral periods or steps

    Braising liquid passes from one batch to the next, so the batches must run in an order that keeps that flow unbroken while meeting demand.

    What is checked

    Liquid flow between batchesWhat must come firstFlow never breaksDemand is met

    How it is scored

    Scored on production timing.

  6. TASK 06

    40-day chip burn-in schedule

    Semiconductor testing Resource-constrained scheduling Design your own approachSeveral periods or steps

    Assign chip burn-in tests to compatible boards, machines and areas over 40 days, within capacity, timing and process rules.

    What is checked

    Output formatResource assignmentCapacity limitsTimingProcess rules

    How it is scored

    Scored on scheduling efficiency and workload pressure, as the project defines them.

  7. TASK 07

    Daily chip burn-in schedule

    Semiconductor testing Resource-constrained scheduling Design your own approachSeveral periods or steps

    Build a day-by-day loading plan for chip burn-in tests over 28 days, reusing boards and machines within capacity, timing and process rules.

    What is checked

    Output formatResource assignmentCapacity limitsTimingProcess rules

    How it is scored

    Scored on scheduling efficiency and workload pressure, as the project defines them.

  8. TASK 08

    Salt drying and packaging schedule

    Salt production Production scheduling Design your own approachSeveral periods or steps

    Plan drying and packaging for orders with due dates, rush orders and changeovers. The plan is replayed step by step to check that resources and stock always add up.

    What is checked

    Production records replayed step by stepResources and stock add upEach step picks up where the last ended

    How it is scored

    Scored on backlog and service level.

  9. TASK 09

    Cutting large rolls to fill orders

    Roll converting Cutting stock Clear-cut model

    Decide which parent rolls and bought-in rolls to cut, and how, so that customer orders are covered and all material is accounted for.

    What is checked

    Allocation formatRight material for each orderDemand is metAll material accounted for

    How it is scored

    Scored on material use and allocation.

  10. TASK 10

    Fixed-ratio coal blending

    Coking Blending Clear-cut model

    Choose a coal blend for coking that stays within stock, price and quality limits.

    What is checked

    Blend formatLimits on each materialStock on handQuality limits

    How it is scored

    Scored on cost, return or quality, as the project specifies.

  11. TASK 11

    Copper-concentrate blending

    Copper smelting Blending Clear-cut model

    Prepare the copper-concentrate batching list for a flash furnace, keeping the mix within composition and stock limits.

    What is checked

    Blend formatLimits on each materialStock on handQuality limits

    How it is scored

    Scored on cost, return or quality, as the project specifies.

  12. TASK 12

    Grouping alloy blocks into smelting batches

    Alloy smelting Batching Clear-cut modelTie-breaking rules

    Divide alloy blocks into smelting batches that meet composition limits and list any blocks left out. Plans are compared on leftovers first, then on a secondary rule.

    What is checked

    Grouping rulesRules within each batchLeftover blocks listed

    How it is scored

    Scored on leftover material first, then on a secondary ordering rule.

  13. TASK 13

    Inventory policy for an existing network

    Supply chain Inventory optimization Design your own approachEstimate from dataSeveral periods or stepsUncertain inputsTie-breaking rules

    Use past data to set finished-goods stock levels for a fixed network of factories and warehouses, using only the data available at each decision date.

    What is checked

    Policy in the required formAllowed parameter valuesService recordsNo use of future data

    How it is scored

    Scored on historical data against several goals, as the project specifies.

  14. TASK 14

    Four-tier supply network

    Supply chain Network design Clear-cut modelDesign your own approachSeveral periods or steps

    Plan how goods flow from suppliers through factories and warehouses to regional warehouses, within capacity and while meeting demand and service targets.

    What is checked

    Valid network linksValid routesFlows balanceCapacity limitsDemand is metService targets

    How it is scored

    Cost and service are recalculated with the project’s accounting rules.

  15. TASK 15

    Supply network with direct shipping

    Supply chain Network design Clear-cut modelDesign your own approachSeveral periods or steps

    Like a regular supply network, except that customers can also be supplied directly. Flows must balance, stay within capacity and meet demand and service targets.

    What is checked

    Valid network linksValid routesFlows balanceCapacity limitsDemand is metService targets

    How it is scored

    Cost and service are recalculated with the project’s accounting rules.

  16. TASK 16

    Supply network with shared central storage

    Supply chain Network design Clear-cut modelDesign your own approachSeveral periods or steps

    Plan a supply network whose central warehouses store several product categories together, within capacity and while meeting demand and service targets.

    What is checked

    Valid network linksValid routesFlows balanceCapacity limitsDemand is metService targets

    How it is scored

    Cost and service are recalculated with the project’s accounting rules.

  17. TASK 17

    One month of refinery production and sales

    Petroleum refining Production planning Clear-cut model

    Plan a month of refinery processing and sales for the highest gross profit. Every input must be accounted for, and products made together must come out in the right proportions.

    What is checked

    Allowed valuesAll material accounted forProcessing recipesJoint-product proportions

    How it is scored

    Revenue and cost are recalculated with the project’s accounting rules.

  18. TASK 18

    Order sequencing on seven plastics lines

    Engineering plastics Parallel-machine scheduling Clear-cut modelSeveral periods or stepsTie-breaking rules

    Assign engineering-plastics orders to seven production lines and put them in order, respecting line capacity and product compatibility, with as few late orders and changeovers as possible.

    What is checked

    Every order assignedLine capacityProduct compatibilityTiming rules

    How it is scored

    Scored on late jobs and product changeovers.

  19. TASK 19

    Picking waves for a warehouse

    Warehousing Order batching Clear-cut modelDesign your own approach

    Group outbound orders into picking waves so that every order is in exactly one wave and no wave is over capacity.

    What is checked

    Wave formatEvery order coveredNo order counted twiceCapacity limits

    How it is scored

    Scored on the quality of the picking waves.

  20. TASK 20

    Long-only index-enhancement strategy

    Asset management Portfolio optimization Design your own approachEstimate from dataSeveral periods or steps

    Write a stock strategy that aims to beat the CSI 2000 index using only the information available on each date. Its trades are run through a fixed backtest.

    What is checked

    Strategy in the required formCode runsTrading rules in the backtest

    How it is scored

    The resulting trades are run through a fixed backtest to recalculate performance.

  21. TASK 21

    Market-neutral long–short strategy

    Asset management Portfolio optimization Design your own approachEstimate from dataSeveral periods or steps

    Write a strategy that buys some stocks and sells others short so that market moves cancel out, within the trading rules. Its trades are run through a fixed backtest.

    What is checked

    Strategy in the required formCode runsTrading rules in the backtest

    How it is scored

    The resulting trades are run through a fixed backtest to recalculate performance.

  22. TASK 22

    Stock portfolio hedged with index futures

    Asset management Portfolio optimization Design your own approachEstimate from dataSeveral periods or steps

    Write a strategy that holds CSI 500 stocks and hedges market risk with IC index futures. Its trades are run through a fixed backtest.

    What is checked

    Strategy in the required formCode runsTrading rules in the backtest

    How it is scored

    The resulting trades are run through a fixed backtest to recalculate performance.

  23. TASK 23

    Three days of grocery delivery

    Grocery delivery Dynamic vehicle routing Design your own approachSeveral periods or stepsUncertain inputs

    Send out vehicles for grocery orders batch by batch as they are confirmed over three days. Dispatches cannot be undone, routes must respect vehicle capacity, and each day’s orders must all go out.

    What is checked

    Dispatch in batchesEvery order coveredValid routesCapacity limitsService targets

    How it is scored

    Scored on delivery performance.

  24. TASK 24

    Hour-by-hour bike rebalancing

    Bike sharing Inventory rebalancing Design your own approachSeveral periods or stepsTie-breaking rules

    Move shared bikes between stations one hour at a time over four hours, as new trip data arrives. Each hour has to start from where the last one left off.

    What is checked

    Plan formatRoutes and stock add upEach step picks up where the last ended

    How it is scored

    The executed plan is replayed and scored on service cost.

  25. TASK 25

    One-hour bike rebalancing

    Bike sharing Inventory rebalancing Clear-cut model

    Plan one hour of moving shared bikes between stations from trip and stock records, scored on the service cost once the plan is carried out.

    What is checked

    Plan formatRoutes and stock add upEach step picks up where the last ended

    How it is scored

    The executed plan is replayed and scored on service cost.

  26. TASK 26

    Bike rebalancing averaged over scenarios

    Bike sharing Stochastic optimization Design your own approachSeveral periods or stepsUncertain inputsTie-breaking rules

    In rolling rounds, choose one four-hour plan for moving bikes that does best on average across four past scenarios, counting both moving costs and lost service.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s rules for execution and scenario risk, with supporting measures.

  27. TASK 27

    Bike rebalancing: decide now, adjust later

    Bike sharing Two-stage stochastic programming Design your own approachSeveral periods or stepsUncertain inputsTie-breaking rules

    Commit to the first hour of bike moves now, then adjust the next three hours separately for each scenario. All scenarios count equally.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s rules for execution and scenario risk, with supporting measures.

  28. TASK 28

    Bike rebalancing: protect key stations

    Bike sharing Chance-constrained optimization Design your own approachEstimate from dataSeveral periods or stepsUncertain inputsTie-breaking rules

    Choose one bike-moving plan for all scenarios, with an extra penalty when key stations, picked from past data, run out of bikes too often.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s rules for execution and scenario risk, with supporting measures.

  29. TASK 29

    Bike rebalancing: limit the worst losses

    Bike sharing Risk-averse optimization Design your own approachSeveral periods or stepsUncertain inputsTie-breaking rules

    Choose one bike-moving plan that balances moving costs, the average loss across scenarios and the average loss in the worst scenarios.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s rules for execution and scenario risk, with supporting measures.

  30. TASK 30

    Bike rebalancing when the odds are uncertain

    Bike sharing Distributionally robust optimization Design your own approachSeveral periods or stepsUncertain inputsTie-breaking rules

    Choose one bike-moving plan that holds up even if the chance of each past scenario is not known exactly, guarding against the worst expected loss within a set range.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s rules for execution and scenario risk, with supporting measures.

  31. TASK 31

    Bike rebalancing: plan for 16 bad cases

    Bike sharing Robust optimization Design your own approachEstimate from dataSeveral periods or stepsUncertain inputsTie-breaking rules

    Build 16 agreed bad cases from past data, then choose one bike-moving plan that keeps the largest loss across them as small as possible.

    What is checked

    Format of each roundOne plan for all scenariosPhysically possibleStock carried over hour to hour

    How it is scored

    Scored by the project’s robust-planning rules and compared with a baseline plan.

What do the skill labels mean?
Clear-cut modelThe business goal is stated and the math model is mostly fixed.
Design your own approachNo method is given: the agent must design its own objective, decomposition, heuristic, policy or rolling scheme.
Estimate from dataDemand, risk or scenario parameters have to be estimated from data.
Several periods or stepsDecisions span several periods or steps, such as batching, sequencing or rolling execution.
Uncertain inputsThe plan has to cope with several scenarios or uncertain inputs.
Tie-breaking rulesPlans that tie, or nearly tie, on the main goal must be ranked by a secondary rule.

A task can carry several labels. They describe the task and play no part in scoring.

Simplified diagram. Click it to open at full size.

Industry
Problem type
How decisions are made
Goal
Skills needed
Required solver

What is checked

    How it is scored

    Task description

    Task files on GitHub