Cover -- Half-title -- Series-title -- Title -- Copyright -- Dedication -- Contents -- Acknowledgements -- Conventions on notation -- Tour d'Horizon -- Whither? Why? -- I Distributional networks -- 1 Simple flows -- 1.1 The setting -- 1.2 Flow optimisation -- 1.3 Seminvariantly scaled costs -- 1.4 A first consideration of optimal design -- 1.5 The dual formulation -- 1.6 The primal and dual forms -- 1.7 The multipliers as marginal prices -- 1.8 Balance of supply and demand -- 2 Continuum formulations -- 2.1 The primal problem -- 2.2 The dual formulation -- 2.3 Evolutionary algorithms -- 2.4 A trivial continuum example -- 2.5 Optimal cooling -- 2.6 More on evolutionary optimisation -- 2.7 Nonuniform spaces -- 3 Multi-commodity and destination-specific flows -- 3.1 Reduction of the problem -- 3.2 The dual and its interpretation -- 3.3 An alternative formulation -- 4 Variable loading -- 4.1 Variable loading and the primal problem -- 4.2 Internal nodes and trunking -- 4.3 Optimality and load splitting -- 4.4 Variable loading in the dual formulation -- 4.5 Degree of trunking -- 4.6 Trunking from a continuous source line -- 4.7 Evolutionary algorithms -- 4.8 Node migration and node splitting -- 4.9 The case of general convex Phi -- 5 Concave costs and hierarchical structure -- 5.1 Incentives to trunking -- 5.2 Other consequences of cost concavity -- 5.3 The combination of environmental penalty and variable load -- 5.4 Hierarchical structure -- 5.5 The outflow function -- 5.6 Optimisation of the trunking rate -- 5.7 The multi-dimensional case -- 6 Road networks -- 6.1 The setting -- 6.2 Structural considerations -- 6.3 Congestion: the queueing analogue -- 6.4 Congestion: fluid and discrete models -- 6.5 When motorists choose -- 7 Structural optimisation: Michell structures -- 7.1 Force and deformation -- stress and strain.

7.2 Braced frameworks: formulation -- 7.3 Reduction of the primal problem -- 7.4 The dual form of the problem -- 7.5 The dual field: Hencky-Prandtl nets -- 7.6 Some examples of Michell structures -- 7.7 The shaping of bone structure -- 8 Computational experience of evolutionary algorithms -- 8.1 Solid materials -- 8.2 Examples -- 8.3 Expanded materials -- 8.4 The literature on structural optimisation -- 9 Structure design for variable load -- 9.1 Return to the coat-hook -- 9.2 Numerical experience -- 9.3 The shaping of bone structure -- 9.4 Buckling -- II Artificial neural networks -- 10 Models and learning -- 10.1 The McCulloch-Pitts net -- 10.2 Back-propagation and Hebb's rule -- 10.3 Linear least-squares approximation -- 11 Some particular nets -- 11.1 Recognition, feedback and memory -- 11.2 The Hamming net -- 11.3 The probability-maximising algorithm -- 11.4 The PMA with compound traces -- 11.5 Comparisons with the olfactory system -- 12 Oscillatory operation -- 12.1 The Freeman model of the neuron -- 12.2 The Freeman oscillator -- 12.3 Oscillation in memory arrays -- III Processing networks -- 13 Queueing networks -- 13.1 The simple queue -- 13.2 The multi-station case -- deterministic treatment -- 13.3 Jackson networks -- 13.4 Optimisation of effort distribution -- 13.5 Queueing networks more generally -- 13.6 A small coda -- 14 Time-sharing processor networks -- 14.1 The fluid and Markov models -- 14.2 The Gittins-Klimov index -- 14.3 Examples -- 14.4 Performance of the index policy -- 14.5 Control with fixed work-station resources -- IV Communication networks -- 15 Loss networks: optimisation and robustness -- 15.1 The linear programming formulation -- 15.2 Design optimisation for a given configuration -- 15.3 A free optimisation -- 16 Loss networks: stochastics and self-regulation -- 16.1 A single exchange -- Erlang's formula.

16.2 Admission control for a single exchange -- 16.3 Equilibrium and asymptotics for the network -- 16.4 Refinements of the asymptotics -- 16.5 Self-regulation for the network -- 17 Operation of the Internet -- 17.1 Some background -- 17.2 A stable regulation rule -- 17.3 An adaptive rate control -- 18 Evolving networks and the Worldwide Web -- 18.1 Random graphs and the Web -- 18.2 Evolution of the scale-free network -- 18.3 Graph properties -- 18.4 Emulation and the power law -- Appendix 1 Spatial integrals for the telephone problem -- A1.1 The limit outflow density -- A1.2 Outflow and withinflow -- A1.3 Character of the outflow density -- A1.4 Flows between cells -- A1.5 Cubic examples -- Appendix 2 Bandit and tax processes -- A2.1 Bandit processes -- A2.2 Tax processes -- A2.3 Adaptation to fixed work-stations -- Appendix 3 Random graphs and polymer models -- A3.1 The zeroth-order model -- A3.2 The first-shell model: unit statistics -- A3.3 Polymer statistics -- A3.4 Nothing but trees -- A3.5 The branching analogue -- Potts criticality -- A3.6 The distribution of degree -- A3.7 Literature and further directions -- References -- Index.