MULTI-OBJECTIVE OPTIMISATION AND THE CASE FOR BESPOKE COMPUTATIONAL DESIGN 

For too long, the conversation around Design for Additive Manufacturing (DfAM) has been dominated by the same tropes, lattices and optimised shapes produced by commercial topology optimisation software. These outputs have become the shorthand for “innovation” in AM, but in most if not all instances, they do not demonstrate functional progress. 

At Metamorphic, we argue that this reliance on off-the-shelf design and optimisation tools has led to a narrowing of the field, not an expansion. Designs are often celebrated for their visual complexity rather than their true performance in real-world conditions. And in advanced design, where geometry should express function, not decoration, most “good-looking” designs fail as they prioritise visual complexity over purposeful performance. True elegance in AM arises when form and function are inseparable, not when form merely imitates intent. 

The future of AM design will not be built on generic algorithms producing approximations of “lightweight” or “optimised” shapes. It will be built on bespoke, intent-driven computational workflows, designed from first principles, optimised across multiple objectives, and tailored to the realities of manufacturability and scale. 

 This is the case for transcending geometry, and designing with intent. 

WHY REDUCTIVE OPTIMISATION LIMITS REAL-WORLD PERFORMANCE 

In additive manufacturing, topology optimisation tools have become indispensable, yet the way they’re often used remains too narrow. 

Designers naturally balance multiple goals (strength, pressure drop, weight, manufacturability), but the computational frameworks they rely on still tend to treat these objectives in isolation. The result is designs that perform beautifully in simulation but struggle in reality. 

A structure that distributes stress perfectly may distort under thermal loads during printing. A design that is only optimised for flow might not be optimised for manufacturing.  Optimisation, in these cases, becomes an exercise in tuning a model, not engineering a part. 

At Metamorphic, we argue that the real challenge isn’t defining more objectives, it’s connecting them. It’s about ensuring that every performance parameter, process constraint, and functional requirement informs a single, coherent design logic. 

Because geometry is a record of intent. 

MULTI-OBJECTIVE OPTIMISATION — WHEN IT SERVES THE INTENT 

At Metamorphic, we recognise that performance can’t always be captured by a single metric, but we also know that optimisation isn’t a one-size-fits-all exercise. 

Where appropriate, we employ multi-objective optimisation to explore geometries across multiple, interdependent criteria such as balancing stiffness with vibrational performance, mass reduction with thermal stability, and heat transfer with pressure drop. 

Every additional optimisation objective tightens the design space and defines its boundaries. Multi-objective optimisation helps us understand the trade-offs and find regions of interest within the design space that satisfy (at varying degrees) our requirements. 

For complex or large domains, we focus on intelligently coupling the parameters that matter most, guided by engineering logic, not algorithmic ambition. 

The goal isn’t to generate endless possibilities, but to identify the variants that matter, the ones that best navigate the trade-offs between intent, performance, and manufacturability. And from there, we scrutinise the highest-scoring candidates with real engineering judgement. 

Computational design was never just about shape. In AM, it becomes a framework for engineering intelligence, turning geometry into a carrier of intent and performance. 

THE CASE FOR BESPOKE COMPUTATIONAL WORKFLOWS 

Why not rely on commercial DfAM software? Because these tools are inherently generic, built to satisfy the broadest user base, not the unique challenges of a specific application. And in practice, most engineers aren’t even starting from a clean slate; they’re handed imported CAD models shaped by the rules of machining, casting, or other types of fabrication. By the time optimisation tools are applied, the design space has already been constrained by assumptions from other manufacturing techniques. 

Rather than expanding creativity, most off-the-shelf DfAM tools end up standardising it. Wrapped neatly within CAD interfaces, they offer preset lattices, cookbook-style recipes, and topology optimisation that reduces mass until constraints are met, not tools that help re-imagine what the component could be. 

This is why so many AM parts look familiar despite being “optimised”, and why when analysing how a part looks it is possible to tell what CAD package it was designed in. They were never redesigned from first principles. They were merely digitally adjusted versions of legacy thinking. 

At Metamorphic, we build bespoke computational workflows tailored to each project. This means writing custom scripts and embedding knowledge of process constraints (whether AM, or casting, for example) into the design loop from the outset. 

For example, in components for optics or semiconductor positioning systems, our focus extends beyond structural performance to include mechanical stiffness and compliance, as well as understanding tolerancing and machining. In energy-sector applications, we balance thermal performance with pressure drop and fluid manifolding with surface roughness considerations. 

This level of customisation ensures that geometry is not a “best guess” from a commercial software package but an intent-driven solution, engineered for both performance and manufacturability. 

DESIGNING WITH MANUFACTURABILITY IN MIND 

One of the greatest flaws in current DfAM practice is the separation of design and manufacturing. Too often, designers treat manufacturability as a downstream problem to be solved after the part has been “optimised.” 

We reject this. At Metamorphic, manufacturability is not a constraint applied at the end, it is a design driver from the start. This includes: 

• Accounting for distortion, residual stress, tolerance drift, and depowdering and cleaning for complex internal geometries during design. 
• Designing self-supporting features and optimising wall thicknesses for both AM and secondary processes. 
• Embedding post-processing foresight into our computational  design scripts, enabling us to visualise machining operations (tool access, fixturing, datum strategies, and material allowances) before production begins, accelerating iteration and eliminating downstream surprises.  

This closed-loop approach ensures that the designs we deliver are not just theoretically optimal but practically realisable and scalable. It is the difference between a neat academic exercise and an industrial solution. 

COMPLEXITY WITH PURPOSE 

In additive manufacturing, complexity is often mistaken for sophistication. But at Metamorphic, we view geometry as both equation and expression. The organic, algorithmic forms we create aren’t complex for their own sake, they’re the visual language of function. When every curve and strut carries meaning, performance and aesthetics become inseparable. 

At Metamorphic, we insist on complexity with purpose. Every feature must earn its place. A braided geometry is not art, it is a solution to fluid mixing or heat exchange. A lattice is not decorative, it is tuned for stiffness-to-weight ratio, thermal conduction, or acoustic damping. 

When complexity is intent-driven, it delivers performance advantages that cannot be matched by conventional methods. And when those geometries are designed with manufacturability in mind, they can scale, from tens of AM parts to thousands of cast components. 

This is where AM stops being an experimental showcase and becomes a backbone of production. 

SCALING INNOVATION 

Even the best AM designs run into scaling challenges. Print speeds, costs, and certification hurdles limit the ability of AM to deliver at volume. That is why our philosophy is hybrid at its core. 

We design geometries that exploit AM’s freedom while remaining transferable to traditional processes like investment casting. This approach provides the best of both worlds, AM’s ability to generate high-performance, functionally novel geometries, and traditional manufacturing’s ability to deliver them at scale, with established materials and qualification routes. 

Scalability is not achieved by buying more AM machines. It is achieved by designing parts that perform at low volume in AM, and continue to perform when manufactured by processes capable of producing thousands per day. 

REDEFING INTENT IN ENGINEERING 

The industry has reached a turning point. The era of celebrating lattices and bone-like structures as symbols of progress must give way to a more rigorous, intent-driven approach. DfAM must evolve from decorative optimisation to strategic engineering, where design intent, material behaviour, and process realities are treated as interdependent, not sequential. True progress is found in integrating human insight with computational intelligence to deliver meaningful, manufacturable performance. 

At Metamorphic, we believe this shift represents the true maturity of additive manufacturing. It is not about replacing traditional methods, nor about treating AM as a gimmick for design showpieces. It is about redefining the relationship between design, process, and performance. 

Beyond topology lies a new paradigm, bespoke computational design, complexity with purpose, and manufacturability as intent. 

SUMMARY 

The future of DfAM will not be decided by teams producing the most eye-catching geometries, but by those who deliver parts that actually work, functionally, certifiably, and at scale. 

At Metamorphic AM, we champion this shift. By combining engineering judgement with bespoke computational design, we show what happens when DfAM escapes the limits of preset tools and inherited CAD assumptions. 

The real question for the industry isn’t how fast we can optimise, it’s whether we’re optimising the right thing in the first place. Too many workflows encourage designers to tune parameters inside pre-defined constraints, creating the illusion of progress without ever revisiting the underlying engineering logic. 

But meaningful innovation in AM doesn’t come from squeezing more performance out of a legacy model. It comes from reframing the problem itself. 

The future isn’t “more objectives” or “more automations”. It’s intent-driven design informed by process behaviour, material realities, and manufacturability from the outset.  

Written By – Laurence Coles and Manolis Papastavrou, Co-Founders Metamorphic AM 

 www.metamorphic.am

The post MULTI-OBJECTIVE OPTIMISATION AND THE CASE FOR BESPOKE COMPUTATIONAL DESIGN  appeared first on Machine Insider.