Skip to main content

Earthmoving Optimisation

Mass-Haul LP · Cut-Fill Balance · Transportation Problem

Construction · Logistics & Flow · Operational

Every highway, dam, railway, and canal starts with moving dirt: excavate where the alignment cuts into a hill, dump where it fills a valley, and truck the difference either from external borrow pits or to external spoil areas. The decision — how much material to haul from each cut station to each fill station, and when to use borrow or spoil — is the classical transportation linear programme. Bruckner drew the first mass-haul diagram in the 1870s; Easa (1988) cast it as a network LP; Moselhi & Alshibani (2008) extended it to mixed material types with compaction factors. The LP is polynomially solvable and scales to projects with hundreds of stations.

Where This Decision Fits

Linear works earthwork planning — the highlighted step is what this page optimises

Alignment SurveyStations + elevations
Cut / Fill TakeoffVolumes per station
Mass-Haul LPOptimal hauls
Equipment & FleetDump-truck routing
ExecuteMonitor progress

The Problem

Balance cut and fill along a linear alignment at minimum haul cost

A linear works project has stations \( s = 1, \ldots, S \) along the alignment. At each station, the engineering design specifies a cut volume \( C_s \ge 0 \) (excavation, excess material available) and a fill volume \( F_s \ge 0 \) (embankment, material required). In addition, there are external borrow pits \( B_j \) (suppliers) and spoil areas \( E_k \) (dumps) with their own costs and capacities. The haul cost per cubic metre from source to destination is \( c_{st} \), typically proportional to distance.

Decision variable \( x_{st} \ge 0 \) is the volume hauled from source \( s \) to destination \( t \).

Mass-haul transportation LP $$ \min \; \sum_{s \in \mathcal{S}^+} \; \sum_{t \in \mathcal{S}^-} c_{st}\, x_{st} $$ $$ \text{s.t.} \quad \sum_{t} x_{st} \;\le\; C_s \qquad \forall s \in \mathcal{S}^+ \quad \text{(supply at cut station/borrow pit)} $$ $$ \sum_{s} x_{st} \;=\; F_t \qquad \forall t \in \mathcal{S}^- \quad \text{(demand at fill station)} $$ $$ x_{st} \;\ge\; 0 \qquad \forall s, t $$

Here \( \mathcal{S}^+ = \{\text{cut stations}\} \cup \{\text{borrow pits}\} \) is the set of sources and \( \mathcal{S}^- = \{\text{fill stations}\} \cup \{\text{spoil areas}\} \) is the set of sinks. Any cut material not hauled to fill is automatically sent to spoil (at positive spoil cost); any unmet fill demand is covered by borrow (at positive borrow cost).

This is a classical transportation problem (Hitchcock 1941, Koopmans 1949) — polynomial-time solvable by the transportation simplex, the network-simplex method, or cost-scaling algorithms. For the typical construction instance with \( S \le 200 \) stations the LP solves in milliseconds. The Bruckner mass-haul diagram (the cumulative plot of \( \sum_{s' \le s} (C_{s'} - F_{s'}) \) against station \( s \)) gives a graphical optimum: minimum-cost haul boundaries correspond to horizontal lines that balance cut and fill within each haul zone.

See Equipment Routing — the CVRP that dispatches the actual dump trucks

Try It Yourself

Solve the mass-haul LP for a linear alignment

Mass-Haul Solver

20 stations
Greedy nearest-match (heuristic)
Northwest-corner + stepping-stone (LP)

The cumulative (cut − fill) curve. Peaks above zero → excess (becomes spoil if long haul uneconomic). Valleys below zero → deficit (filled from borrow or reverse haul).

Pick a scenario and click Solve.

The Algorithms

From Bruckner's diagram to transportation simplex to network flow

Graphical

Bruckner Mass-Haul Diagram (1870s)

Manual, O(S) computation of cumulative volumes

Plot the cumulative \( M_s = \sum_{s' \le s} (C_{s'} - F_{s'}) \) against station. Draw horizontal "balance lines" at various levels; the lowest horizontal line that leaves the curve above it everywhere between two stations defines a haul zone. Within each zone, all cut material is used by nearby fill — no external borrow or spoil needed. Pedagogically beautiful and still drawn on every civil-engineering exam.

Heuristic

Greedy Nearest-Match

O(S\u00b2)

For each cut station (in some order), ship material to the nearest fill station that still has unmet demand. Very simple; the default algorithm on this page. Typically produces cost within 15–25% of LP optimum on real alignments; poor on tricky geometries where reverse hauls are needed.

Exact (LP)

Transportation Simplex

Polynomial — network simplex is O(V\u00b2 E log(V C))

Start with a basic feasible solution from the northwest-corner rule (or Vogel's approximation), then apply stepping-stone cost improvements: identify an entering non-basic variable with negative reduced cost, compute the cycle, pivot. Provably finds the LP optimum. Easa (1988) wrote the construction-specific LP; Jayawardane & Harris (1990) refined it for multi-material earthworks. The second algorithm on this page.

Real-World Complexity

Why real earthmoving goes beyond textbook mass-haul LP

Beyond the Basic Transportation LP

  • Material types — Not all cut is usable as fill. Rock, common earth, organic soils, and clays have different engineering properties; fill at a station may demand a specific material. Moselhi & Alshibani (2008)'s multi-material MIP handles this.
  • Compaction (shrinkage / swell) factors — 1 m\u00b3 of in-situ earth becomes 1.25–1.35 m\u00b3 once loaded (swell) and 0.85–0.95 m\u00b3 once compacted. Cut and fill volumes live on different scales; the LP uses bank/loose/compacted cubic metres explicitly.
  • Grade / access restrictions — Some hauls are blocked by intermediate features (bridges, creeks, bog). The network graph is partial, not complete.
  • Temporary stockpiles — Cut may be stockpiled mid-project if fill isn't yet ready. This turns the LP into a multi-period dynamic transportation problem.
  • Haul-road cost — Cost per m\u00b3 depends on whether there's a prepared haul road, whether trucks can run at full speed, weather. Dynamic re-optimisation applies.
  • Equipment productivity — The LP gives volumes; actual execution requires fleet sizing, cycle-time calculation, and queueing analysis (AbouRizk & Hajjar 1998's simulation models).
  • Environmental regulation — Spoil-area permits, erosion-control requirements, and haul-route noise limits add constraints beyond cost.

Related Logistics & Flow Variants

Mass-haul is the first-class transportation LP in construction OR

Equipment Routing (CVRP). Once the LP gives volumes, dump-truck routing dispatches the fleet. Mass-haul LP feeds quantities into CVRP. → Open Equipment Routing
Linear / LOB Scheduling. Earthmoving in highway projects is the first repetitive activity; LSM and mass-haul together plan the linear earthwork. → Open Linear Scheduling
Site Layout Planning. For point projects (dam, airport), temporary stockpile placement is an SLP sub-problem tied to mass-haul volumes. → Open Site Layout
Material Procurement. Borrow-pit contracts, haul-road construction, and dump-permit planning are upstream procurement decisions that feed into mass-haul costs. → Open Material Procurement
RCPSP. Earthmoving is itself a set of activities (strip topsoil, excavate rock, place fill, compact) with precedence and equipment-pool constraints. → Open Project Scheduling
Dynamic / Multi-material Mass-Haul. Moselhi-Alshibani 2008 and Son-Hendrickson 2005 extend the LP with material types, compaction, and time phases; active research.

Key References

Cited above · DOIs & permanent URLs

Easa, S. M. (1988).
“Earthwork allocations with nonconstant unit costs.”
Journal of Construction Engineering and Management, 114(4), 641–655. (The canonical transportation-LP formulation for construction mass-haul.) doi:10.1061/(ASCE)0733-9364(1988)114:4(641)
Moselhi, O., & Alshibani, A. (2008).
“Optimization of earthmoving operations in heavy civil engineering projects.”
Journal of Construction Engineering and Management, 135(10), 948–954. (MIP with multi-material, compaction, and equipment constraints.) doi:10.1061/(ASCE)CO.1943-7862.0000059
Marzouk, M., & Moselhi, O. (2004).
“Multiobjective optimization of earthmoving operations.”
Journal of Construction Engineering and Management, 130(1), 105–113. (Cost + time + quality multi-objective GA for earthmoving.) doi:10.1061/(ASCE)0733-9364(2004)130:1(105)
Jayawardane, A. K. W., & Harris, F. C. (1990).
“Further development of integer programming in earthwork optimization.”
Journal of Construction Engineering and Management, 116(1), 18–34. (Extension of Easa's LP to integer-programming cases; multi-material handling.) doi:10.1061/(ASCE)0733-9364(1990)116:1(18)
Son, J., Mattila, K. G., & Myers, D. S. (2005).
“Determination of haul distance and direction in mass excavation.”
Journal of Construction Engineering and Management, 131(3), 302–309. (Bruckner-diagram revisited; haul-direction heuristics.) doi:10.1061/(ASCE)0733-9364(2005)131:3(302)
Hitchcock, F. L. (1941).
“The distribution of a product from several sources to numerous localities.”
Journal of Mathematics and Physics, 20(1-4), 224–230. (The original transportation problem formulation.) doi:10.1002/sapm1941201224
Koopmans, T. C. (1949).
“Optimum utilization of the transportation system.”
Econometrica, 17, 136–146. (Transportation LP and its economic interpretation.) doi:10.2307/1907301
AbouRizk, S., & Hajjar, D. (1998).
“A framework for applying simulation in construction.”
Canadian Journal of Civil Engineering, 25(3), 604–617. (Simulation complements to mass-haul LP for fleet sizing and productivity.) doi:10.1139/l97-122
Kim, B., & Russell, J. S. (2003).
“Framework for an intelligent earthwork system.”
Automation in Construction, 12(1), 1–14. (GPS + mass-haul integration for automated earthwork.) doi:10.1016/S0926-5805(02)00032-X

Earthmoving a highway, dam, or major civil project?

From mass-haul LP to fleet sizing and GPS-integrated automated earthwork, OR turns dirt-moving from rule-of-thumb into optimised logistics. Let's discuss how Operations Research can sharpen your earthworks.

Disclaimer
Cut/fill volumes, haul costs, and borrow/spoil capacities are illustrative values. Real earthmoving requires material classification, compaction factors, access constraints, weather windows, and equipment cycle times not captured here.
Back