Skip to content

Deterministic models

Eight chapters where every datum is known: you decide under certainty, and the value lies in interrogating the solution (how much is one more unit of a resource worth? where does the plan break?).

Inside you will find QPs, convex NLPs and a non-convex NLP: the modelling language changes, the workflow does not --- build the model, solve it, read the multipliers, move the data and see what holds.

  • Production with quadratic smoothing


    The quadratic variant of the production case: a penalty on the jumps levels the plan; QP multipliers and the cost-smoothness frontier.

    Convex QP

  • Supply chain and CO₂


    Minimum-cost flow, convex congestion and the cost-emissions frontier with an internal carbon price.

    Convex NLP

  • Markowitz portfolio


    The most famous QP in history: efficient frontier, diversification and the fragility of the estimates.

    QP

  • Pricing and revenue management


    Endogenous demand, bilinear objective, corner optima and the true value of one more seat.

    NLP

  • Advertising budget


    Diminishing returns, the KKT condition «equal marginal return on every active channel», value-versus-budget curve.

    Convex NLP

  • Continuous location


    Centroid, Weber point and minimax on the map: efficiency versus equity, with a geographic constraint.

    Convex NLP

  • EV charging with a smooth profile


    The quadratic variant of overnight charging: the total draw is smoothed, the multipliers of the vehicles coincide.

    Convex QP

  • Queues and service capacity


    The utilisation wall: optimal capacity, the price of a service promise, robustness to uncertain demand.

    Convex NLP