Overview | Planning formulation | Landscape and ecosystem-service data | Define the restoration decision context | Build the planning problem | Express restoration effects as deficits | Register five restoration objectives | Generate alternatives with AUGMECON | Examine the many-objective solution set | Run design and objective values | Read the solution set as a sequence of decisions | How much deficit targeting accompanies additional cost? | Compare the complete performance profile of each plan | Identify extremes and representative compromises | Connect objective performance to spatial decisions | Compare decision-space similarity | Interpretation and limitations

1. Start from one spatial landscape | 2. Name combinations with action sets | 3. Change the number of actions allowed in each unit | 4. Specify dependencies and compatibility | Directional requirements and alternatives | Exclusions and selection together | 5. Compare three spatial prescriptions | 6. Distinguish selecting a combination from describing its joint effect | 7. Practical rules and common pitfalls

Preserve inputs and record assumptions | Exact activation and model size | Solve with all three MO methods | Negative interactions, net benefit, and final loss | Targets on joint outcomes

Build a common base | Compare solver settings and MO methods | Accumulate independent constraints | Accumulate compatible locks

Overview | The MaPP Rwanda training case | Define the reserve problem | Minimum cost and the boundary length modifier | Going beyond a fixed BLM | Compare representative reserve designs | Interpretation | References

The decision comes before the method | Choosing a method | Preferences before or after optimisation | Quick choice | The methods | Weighted sum: state relative preferences | Decision question | Main idea | Use weighted sum when | Use caution when | Epsilon-constraint: state acceptable limits | Use epsilon-constraint when | AUGMECON: construct a cleaner efficient frontier | Use AUGMECON when | A practical workflow | 1. Establish objective ranges | 2. Decide how preferences can be expressed | 3. Start coarse | 4. Inspect both objective values and spatial plans | 5. Document the decision process | Mathematical background | A common mathematical view | Pareto efficiency | What does lexicographic = TRUE mean? | Summary

1) Calculating benefits when threats are discrete (presence/absence) | Not proportional probability of persistence | 2) Calculating benefits when threats are continuous | Example with continuous intensities of threats | References

Overview | Workflow | Usage | Planning units data | Features data | Threats data | Boundary data | Step 1: Initialize the problem | Step 2: Create the mathematical model | Step 3: Solve the model | Getting information about solutions | getActions() | getSolutionBenefit() | getCost() | getConnectivityPenalty() | getPerformance() | Sensitivity analyses on blm, budget and target parameters | References

1) Preparing and analyzing input data | 2) Base model | 3) Model with different curve param | 4) Model with connectivity requirements | 5) Conclusions

1) Model I: Only recovery targets | 2) Model II: Recovery and conservation targets | 3) Model III: Only recovery targets and connectivity | 4) Model IV: Recovery and conservation targets and connectivity

Experimental settings | Gurobi | CPLEX | CBC | Symphony | Results | Academic solvers: gurobi vs CPLEX | Non-academic solvers: CBC vs symphony | Conclusions