Question 11 of 27

Question 2.3 truck dispatch systems (7 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

2.3.1 — dispatch algorithm families. Best Path (BP) is a simple, myopic heuristic (e.g. always send the next available truck to whichever shovel currently has the shortest queue, or the highest match-factor deficit) — fast and easy to implement, but only locally optimal, one decision at a time. Linear Programming (LP) formulates truck-to-shovel assignment as an optimisation problem (minimise total travel/waiting time or maximise ore+waste throughput subject to blend/grade and fleet-capacity constraints) solved simultaneously for the WHOLE fleet each dispatch cycle — globally optimal for that snapshot, but assumes the input data (queue lengths, cycle times) are static for the solve interval. Dynamic Programming (DP) extends this over a rolling time horizon, evaluating sequences of assignment decisions to find the path that is best not just now but over several look-ahead cycles — more computationally demanding but captures how today’s assignment affects tomorrow’s queue state.

2.3.2 — the dispatch system. Modern systems (e.g. Modular Mining DISPATCH, Wenco) run an LP/assignment algorithm (often a transportation-problem or network-flow LP, re-solved every time a truck becomes available or every few minutes) on a central server, communicating assignments over a wireless (radio/WiFi/LTE) network to an in-cab computer/tablet with GPS on every truck and shovel. Trucks report back to the server: GPS position, payload (via an on-board weighing/suspension-load system), cycle-state (loading/hauling/dumping/queueing), and fuel/engine data. The server sends each truck: its next assigned loading unit and dump/crusher destination, plus route guidance. Shovels report their own queue state and, on hydraulic/rope shovels, bucket-pass payload data used for grade reconciliation.

2.3.3 — how planning targets are built in. Maximum production is encoded as the LP’s primary objective function (maximise total tonnes moved per period, subject to truck/shovel availability). Stripping ratio is encoded as a CONSTRAINT (or a weighted secondary objective) forcing the dispatch algorithm to allocate a minimum share of truck-hours to waste even when ore alone would maximise short-term tonnes, so the plan’s waste:ore ratio target is not silently abandoned for short-term truck efficiency. Grade control is encoded as a blending constraint at the crusher/stockpile destination — the dispatcher routes ore trucks to whichever destination (primary crusher feed, low-grade stockpile, blend pad) keeps the DELIVERED head grade within the mill’s target band, overriding the pure shortest-path assignment when grade compliance requires it. All three run inside the SAME re-solved LP/DP cycle (frequently, every few minutes) so priorities are continuously rebalanced as queues, grades and stripping progress actually evolve through the shift.