You are currently browsing the category archive for the ‘networks’ category.
Volume 42 of the AER, published in 1952, contains an article by Paul Samuelson entitled `Spatial Price Equilibrium and Linear Programming’. In it, Samuelson uses a model of Enke (1951) as a vehicle to introduce the usefulness of linear programming techniques to Economists. The second paragraph of the paper is as follows:
In recent years economists have begun to hear about a new type of theory called linear programming. Developed by such mathematicians as G. B. Dantzig, J. v. Neumann, A. W. Tucker, and G. W. Brown, and by such economists as R. Dorfman, T. C. Koopmans, W. Leontief, and others, this field admirably illustrates the failure of marginal equalization as a rule for defining equilibrium. A number of books and articles on this subject are beginning to appear. It is the modest purpose of the following discussion to present a classical economics problem which illustrates many of the characteristics of linear programming. However, the problem is of economic interest for its own sake and because of its ancient heritage.
Of interest are the 5 reasons that Samuelson gives for why readers of the AER should care.

This viewpoint might aid in the choice of convergent numerical iterations to a solution.

From the extensive theory of maxima, it enables us immediately to evaluate the sign of various comparativestatics changes. (E.g., an increase in net supply at any point can never in a stable system decrease the region’s exports.)

By establishing an equivalence between the Enke problem and a maximum problem, we may be able to use the known electric devices for solving the former to solve still other maximum problems, and perhaps some of the linear programming type.

The maximum problem under consideration is of interest because of its unusual type: it involves in an essential way such nonanalytic functions as absolute value of X, which has a discontinuous derivative and a corner; this makes it different from the conventionally studied types and somewhat similar to the inequality problems met with in linear programming.

Finally, there is general methodological and mathematical interest in the question of the conditions under which a given equilibrium problem can be significantly related to a maximum or minimum problem.
This is the second posts about stability and equilibrium in trading networks. The first post may be found here. In it I assumed a single homogenous divisble good being traded by agents with quasilinear preferences. Each agent was associated with a vertex of a graph and the edges represented pairs of agents who were permitted to trade with each other. Furthermore, each agent was designated the role of buyer and seller. Buyers could only trade with sellers and viceversa. This meant the underlying graph was bipartite. In this post I will drop the assumption that they trade a homogenous good and that it is dvisible.
As noted in the earlier post, when the utility of buyers and the costs of sellers are linear in quantity, total unimodularity of the underlying constraint matrix is sufficient to ensure the existence of Walrasian prices that support an efficient allocation that is integral. If buyer utility functions become concave or sellers’s cost functions become convex, this is no longer true.
Suppose instead that the buyers utility functions are Mconcave and seller’s cost functions are Mconvex (object to be defined soon), then, the efficient allocation (the precise formulation will appear below) is indeed integer valued and supporting Walrasian prices exist. The corresponding allocations and prices can be interpreted as stable outcomes which cannot be blocked. This result was established by Murota and Tamura (2001).
I’ll define Mconcavity. For any integer valued vector in let and . A real valued function defined on the integer vectors of (with nonempty domain) is M concave if for all in its domain and any index there exists an index in such that
Here is the 01 vector with a zero in every component except component . I describe three different ways one might interpret this notion.
First, one can think of Mconcavity as trying to capture an essential feature of the basis exchange property of matroids. If you don’t know what a matroid is, skip to the next paragraph. Think of the vectors and as being 01 and representing subsets of columns of some given matrix, , say. Define to be the rank of the columns in the subset associated with . What the definition of Mconcavity captures is this: for any column, I remove from , that is not in and add it to , I can find (possibly) a column from (not in ) to add to and not diminish the sum of the ranks of the two sets. The argument is straightforward.
A second interpretation is to be had by comparing the definition of Mconcavity to the following way of stating concavity of a function:
If and were numbers with , then, the first term on the right represents a step from closer to and the second term is a step from closer to . Mconcavity can be interpreted as a discrete analogy of this. Think of being above and to the left of . A step closer to would means one step down in the direction and a step to the right in the direction., i.e. . The vector has a similar interpretation.
The third interpretation comes from an equivalent definition of Mconcavity, one that shows that it is identical to the notion of gross substitutes proposed by Kelso and Crawford (1982). Let be a nonnegative price vector. Let be a nonnegative integer that maximizes , i.e., a utility maximizing bundle. Consider a new price vector . Then there exists a new utility maximizing bundle such that for all such that . In words, the demand for goods whose price did not change cannot go down.
The key result of Murota and Tamura is that the following linear program has an integer optimal solution when the ‘s are Mconcave and the ‘s are Mconvex.
subject to
Notice, that buyer’s utilities do not necessarily depend on the total amount consumed. They can depend on the vector of inflows from the various sellers (similarly for the sellers). We can interpret this to mean that the sellers sell different goods which are substitutes for each other and there is an upper limit, on the amount of seller ‘s good that can go to seller . Such a bound is necessary to make sure the problem is well defined. Consider the case when utilities and costs are linear in quantity. The problem could be unbounded then. A second interpretation is that seller’s sell the same good but under different contractual terms (payment plan, delivery date etc.) which make them, in the buyer’s eyes, imperfect substitutes. The only element of the contract not fixed is the price and that is what is determined in equilibrium.
This is the first of a series of posts about stability and equilibrium in trading networks. I will review and recall established results from network flows and point out how they immediately yield results about equilibria, stability and the core of matching markets with quasilinear utility. It presumes familiarity with optimization and the recent spate of papers on matchings with contracts.
The simplest trading network one might imagine would involve buyers () and sellers () of a homogenous good and a set of edges between them. No edges between sellers and no edges between buyers. The absence of an edge in linking and means that and cannot trade directly. Suppose buyer has a constant marginal value of upto some amount and zero thereafter. Seller has a constant marginal cost of upto some capacity and infinity thereafter.
Under the quasilinear assumption, the problem of finding the efficient set of trades to execute can be formulated as a linear program. Let for denote the amount of the good purchased by buyer from seller . Then, the following program identifies the efficient allocation:
subject to
This is, of course, an instance of the (discrete) transportation problem. The general version of the transportation problem can be obtained by replacing each coefficient of the objective function by arbitrary numbers . This version of the transportation problem is credited to the mathematician F. J. Hitchcock and published in 1941. Hitchcock’s most famous student is Claude Shannon.
The `continuous’ version of the transportation problem was formulated by Gaspard Monge and described in his 1781 paper on the subject. His problem was to split two equally large volumes (representing the initial location and the final location of the earth to be shipped) into infinitely many small particles and then `match them with each other so that the sum of the products of the lengths of the paths used by the particles and the volume of the particles is minimized. The ‘s in Monge’s problem have a property since called the Monge property, that is the same as submodularity/supermodularity. This paper describes the property and some of its algorithmic implications. Monge’s formulation was subsequently picked up by Kantorovich and the study of it blossomed into the specialty now called optimal transport with applications to PDEs and concentration of measure. That is not the thread I will follow here.
Returning to the Hitchcock, or rather discrete, formulation of the transportation problem let be the dual variables associated with the first set of constraints (the supply side) and the dual variables associated with the second or demand set of constraints. The dual is
subject to
We can interpret as the unit price of the good sourced from seller and as the surplus that buyer will enjoy at prices . Three things are immediate from the duality theorem, complementary slackness and dual feasibility.
 If is a solution to the primal and an optimal solution to the dual, then, the pair form a Walrasian equilibrium.
 The set of optimal dual prices, i.e., Walrasian prices live in a lattice.
 The dual is a (compact) representation of the TU (transferable utility) core of the cooperative game associated with this economy.
 Suppose the only bilateral contracts we allow between buyer and seller are when . Furthermore, a contract can specify only a quantity to be shipped and price to be paid. Then, we can interpret the set of optimal primal and dual solutions to be the set of contracts that cannot be blocked (suitably defined) by any buyer seller pair .
 Because the constraint matrix of the transportation problem is totally unimodular, the previous statements hold even if the goods are indivisible.
As these are standard, I will not reprove them here. Note also, that none of these conclusions depend upon the particular form of the coefficients in the objective function of the primal. We could replace by where we interpret to be the joint gain gains from trade (per unit) to be shared by buyer and seller .
Now, suppose we replace constant marginal values by increasing concave utility functions, and constant marginal costs by ? The problem of finding the efficient allocation becomes:
subject to
This is an instance of a concave flow problem. The KuhnTuckerKarush conditions yield the following:
 If is a solution to the primal and an optimal Lagrangean, then, the pair form a Walrasian equilibrium.
 The set of optimal Lagrange prices, i.e., Walrasian prices live in a lattice.
 Suppose the only bilateral contracts we allow between buyer and seller are when . Furthermore, a contract can specify only a quantity to be shipped and price to be paid. Then, we can interpret the set of optimal primal and dual solutions to be the set of contracts that cannot be blocked (suitably defined) by any buyer seller pair .
Notice, we lose the extension to indivisibility. As the objective function in the primal is now concave, an optimal solution to the primal may occur in the interior of the feasible region rather than at an extreme point. To recover `integrality’ we need to impose a stronger condition on and , specifically, they be concave and convex respectively. This is a condition tied closely to the gross substitutes condition. More on this in a subsequent post.
Recent Comments