Optimisation searches for the best feasible allocation or schedule against a stated objective. Time, capacity, service promises, labour rules and cost often compete, so an intuitive plan can waste resources or fail an important constraint. Marketways translates the business outcome and operating limits into the model, then tests whether the solution remains practical. The result is a plan that can be implemented and explained rather than a mathematically neat answer detached from operations.
The decision this method supports
We use Optimisation & Scheduling to help clients answer: Which feasible action performs best against the objective and operating limits?
How the method works
Optimisation searches for the best feasible choice against an explicit objective. Scheduling applies this reasoning to the timing and assignment of work. The method requires clear constraints because an answer that ignores capacity, labour rules or service promises is not operationally useful.
A business example
A delivery business may minimise travel time while respecting vehicle capacity, delivery windows and driver hours. The resulting plan is useful only if the objective also reflects customer priorities and the exceptions dispatchers must manage.
How the client uses the result
Optimisation helps a business use scarce time, people, assets and money more effectively. It can compare many feasible combinations and identify an allocation or schedule that performs best against the business objective and operating constraints.
What we deliver
We produce forecast ranges, scenarios, risk distributions, prioritised alternatives or an optimised plan. A manager should understand which assumptions drive the result and what conditions would trigger a different action.
Limits and complementary methods
Optimisation finds the best answer to the stated problem. A poorly chosen objective or missing constraint can make the mathematically best answer harmful in practice.
Selected methods and techniques
We select from these established methods according to the decision, evidence and operating conditions.
- Assignment optimisation: Map required tasks, routes or shifts to eligible people, vehicles or other resources so an explicit objective is improved while capacity, skills, availability, location and other assignment rules are satisfied. Report work that cannot be assigned and which constraint causes the conflict. Assignment decides which resource handles which work; scheduling decides when it happens, and routing decides the order and travel path between locations.
- Capacitated VRP: Formulate and solve the capacity-constrained variant of the Vehicle Routing Problem for a defined set of stops, depots and vehicles. Each required stop is assigned to a vehicle and placed in a route while the relevant load, weight, volume or other capacity limit is respected and the stated distance, cost or service objective is evaluated. CVRP is the model class; this method is its application to a particular case. Time windows, multiple trips, split deliveries or mixed fleets apply only when the formulation includes them.
- Capacity modelling: Translate demand and service requirements into the work that people, equipment, space and supporting resources can complete over a stated period. Represent processing times, availability, shifts, maintenance, yield, queues, bottlenecks and any shared-resource constraints, then test utilisation and service performance under alternative volumes. Distinguish nameplate or installed capacity from practical capacity after real operating losses. Capacity modelling shows whether resources can serve assumed demand; it does not establish that the demand exists or that serving it will be profitable.
- Heuristics / metaheuristics: Use structured search procedures to find feasible, high-quality solutions when enumerating every possible plan or proving the best one is impractical. A heuristic uses problem-specific construction or improvement rules; a metaheuristic, such as tabu search, simulated annealing or a genetic algorithm, directs a broader search across many candidate solutions. Define constraint handling, starting conditions, stopping rules and any random seeds or repeated runs. Compare results with a baseline, an exact solution on smaller cases or, where available, a bound showing the best result any feasible solution could possibly reach. The best solution found is not automatically the global optimum.
- Integer programming: Formulate an optimisation problem in which the decision variables are required to take integer or binary values, with an explicit objective and stated constraints. In this ontology, use Integer programming for an all-integral formulation; use Mixed-integer programming when the model also contains continuous decisions. Report whether the solver found a feasible plan, the best plan found, any bound on the best possible result and the remaining gap between that bound and the plan, together with time or tolerance limits. Call a solution proven optimal only when the solver status supports that claim.
- Location / site assessment: Define the requirements a location must satisfy and compare candidate sites on customer access, labour, inputs, utilities, logistics, regulation, environmental conditions, capacity, cost, resilience and implementation timing. Apply hard constraints before weighted preferences, verify critical claims through documents or site inspection and test how conclusions change with uncertain demand or costs. The method judges site suitability for a stated operating model. A strong location cannot compensate for an infeasible business, and the cheapest site may not provide the required service or capacity.
- Location analysis: Compare locations or operating footprints within the entry design against the requirements of a defined customer geography and service model. Examine customer and supplier access, travel or delivery time, labour, utilities, regulation, property, capacity, cost, resilience and expansion options; apply hard constraints before preferences; and account for overlapping catchments and network effects. Test alternatives using consistent demand and operating assumptions. Location analysis helps choose where facilities or teams should sit after or alongside market selection. It is not country screening, property valuation or detailed vehicle-routing optimisation, and a low-cost site is not suitable if it cannot meet the required service.
- Maintenance-policy optimisation: Compare feasible inspection, preventive, condition-based, predictive, repair and replacement policies and select the policy or interval that best meets a stated cost, risk, availability or service objective. Represent failure and degradation behaviour, action effectiveness, planned and unplanned downtime, labour, spares, production consequences, resource constraints and uncertainty, and show how the decision changes when material assumptions vary. Reliability-Centred Maintenance supplies a wider consequence-led logic for choosing applicable maintenance responses; maintenance-policy optimisation is the quantitative comparison of specified alternatives.
- Mixed-integer programming: Formulate an optimisation problem that combines integer or binary choices with continuous quantities under one objective and set of constraints. Use it when decisions such as opening a depot, selecting a vehicle or activating a shift are indivisible while quantities such as flow, time or fuel can vary continuously. Report whether the solver found a feasible plan, the best plan found, any bound on the best possible result, the remaining gap between that bound and the plan and the stopping conditions. The best feasible solution found so far is not a proven optimum unless the solver status supports that claim.
- Network-flow optimisation: Represent a connected operating system as nodes and directed links, attach supplies, demands, capacities and costs or travel times, and solve for the quantities that should move through the network. Enforce flow conservation so what enters, leaves, is supplied or is consumed reconciles at each node. The Network Flow Models framework defines the reusable model family; this method constructs, populates and solves a particular instance. The result is an aggregate flow plan and does not automatically specify individual vehicle tours or visit order.
- Queueing analysis: Estimate congestion and service performance when arrivals compete for limited people, vehicles, bays or other service capacity. Specify the arrival and service patterns, queue discipline, number and availability of servers, capacity, priorities, abandonment and relevant time horizon; then estimate waiting time, queue length, utilisation and service-level risk. Check whether steady-state or time-varying assumptions fit the operation and compare material results with observed data. Spare average capacity does not guarantee short waits when arrivals or service times are variable.
- Scheduling optimisation: Choose task start and finish times, sequence and resource use over a defined planning horizon so an explicit objective is improved while precedence, calendars, availability, shifts, breaks, setup times and service windows are respected. Report infeasible, late or unplaced work and the constraint responsible. Scheduling decides when and in what order work happens; assignment decides which resource performs it, and routing adds travel between locations.
Parent method family
Forecasting, Risk & Optimisation explains how this method connects to adjacent methods and relevant services.
Related service families
These service families contain business questions supported by this method. Service pages link to the wider method family so readers can understand the complete analytical approach.